A Geophysical Investigation of the Derwent Estuary
David J Gibbons B. Sc.
UNIVERSITY OF TASMANIA
A research thesis submitted in partial fulfillment of the requirements of the Degree of Bachelor of Science with Honours
School of Earth Sciences, University of Tasmania November, 2001 \.
Acknowledgements
(, Michael Roach, my supervisor and chief guru for your assistance, expertise and support (even if you did reckon the funny bits in the seismic were basalt!). Thanks especially for scraping together the funds for the project after the grant application got rejected. I hope you enjoyed your holiday, you certainly deserved it.
(, James Reid - stand-in guru and all-around good guy - for your help and good humour, particularly in Michael's absence. Thanks also for your lessons in the dark art fortran 77.
Alan Jordan and Miles Lawler from the Tasmanian Aquaculture and Fisheries " Institute, without whom this project could not have proceeded. Alan for providing' 'mates rates' for the vessels and Miles for piloting them back and forth, back and forth, back and forth....thank you both.
David Mitchell from the University of Sydney, for his willingness to come to Hobart /, in the colder months (straight from the North West SheW) to conduct our seismic survey. His good humour and patience with my clumsiness ("Please don't stand on the eel, Dave!") certainly made the seismic survey a more pleasant experience than it might otherwise have been. .. Pat Quilty for his willingness to help whenever required (particularly in terms of my' literature review). Thanks also to Peter Harris, for looking at my seismic data early in the year and allowing me to use his carbonate distribution map in my thesis.
Mineral Resouces Tasmania for their support both personally (through a State Government Mining Scholarship) and for their support of the project. Thanks " particularly to the staff who attended and gave input at the mid-year thinktank, aka 'Roachy's presentation' ("Hi Roachy, we've come for your presentation"). Thanks also to Dr. David Leaman for his attendance and input at 'Roachy's presentation', and for other valuable discussions throughout the year.
.. ,.~ The library staff at the SciTech and Morris Miller libraries here at Uni, the Mineral Resources Tasmania library at Rosny, and the CSIRO Marine library at Salamanca.
Cheers to my fellow honours students and any staff member or undergrad that said anything nice about me or helped me during the year. Special thanks to those who C' offered to review my chapters in the absence of Michael and James during the latter part of the year - Ben Jones, Luisa D'Andrea, Nolene Dorn, Geoff Peters and my brother Geoff Gibbons. Perhaps I should have taken up the offers! Special thanks also to Jodie Cutler and her friend George for their help with my fieldwork.
(', Thanks, Dave.
" II "
Table of Contents
ABSTRACT I (\ ACKNOWLEDGEMENTS II LIST OF FIGURES V LIST OF TABLES VIII
(' CHAPTERl INTRODUCTION 1 THE DERWENT RIVER 1 AIMS 4 TECHNIQUES (FOR MARINE SURVEYS) 5 POTENTIAL FIELD METHODS 5 ELECTRICAL AND ElM 5 C SEISMIC 6 SONAR 6 CHAPTER 2 GEOLOGY 8 INTRODUCTION 8 PREVIOUS WORK 8 PREVIOUS GEOPHYSICAL SURVEYS 9 ( GEOLOGY OF lliE HOBART AREA 12 STRATIGRAPHY LOWER PARMEENER SUPER-GROUP 14 UPPER PARMEENER SUPER-GROUP 14 TERTIARY 18 QUATERNARY 20 ( '., IGNEOUS ROCKS JURASSIC DOLERITE 22 TERTIARY BASALT 24 STRUCTURE AND STRUCTURAL mSTORY 26 PErROPHYSICAL PROPERTIES 26 MAGNETIC PROPERTIES 27 ( , ACOUSTIC PROPERITES 28 CHAPTER 3 MAGNETIC SURVEY 30 FIELDWORK 30 SPACE WEATHER 33 DATA PROCESSING 35 HEADING TEST 43 (' CHAPTER 4 MAGNETIC MODELING 50 INTRODUCTION 50 SENSITIVITY TESTING 51 MODELING 57 CHAPTER 5 SEISMIC SURVEY 71 SEISMIC TERMS 72 ( , ACOUSTIC TURBIDITY 77 THE BASALT THEORY 77 DISPROVING THE BASALT lliEORY 79 lliE EVIDENCE FOR SHALLOW GAS 79 ACOUSTIC TURBIDITY 82 ENHANCED REFLECTIONS 82 (\ ACOUSTIC BLANKING 83 PHASE REVERSALS 83 GAS SEEPS 85 VELOCITY PULLDOWN 85 INDIRECTEVIDENCE 85
(\ III (' BIOGENIC METHANE PRODUCTION 92 SEDIMENT SAMPLING 96 OTHER SEISMIC FEATURES 100 CHAPTER 6 ACOUSTIC FINITE DIFFERENCING 106 CONCLUSIONS 113
(\ CHAPTER 7 SEISMIC INTERPRETATION 114 CHAPTER 8 BAT~ETRICSURVEY 157 CHAPTER 9 COMBINED GEOLOGICAL INTERPRETATION 162 CHAPTER 10 CONCLUSIONS AND FURTHER WORK 168 (, CONCLUSIONS 168 RECOMMENDATIONS FOR FURTHER WORK 168
REFERENCES
APPENDIX 1 LITERATURE REVIEW 'GLOBAL SEA LEVEL CHANGE OVER THE LAST I' 250,000 YEARS' APPENDIX 2 SOURCE CODE FOR FORTRAN 77 PROGRAM 'HEADING' (WRITTEN BY D. GffiBONS) APPENDIX 3 SOURCE CODE FOR FORTRAN 77 PROGRAM 'GE02UTM' (WRITTEN BY D. GffiBONS) APPENDIX 4 SOURCE CODE FOR QUICKBASIC PROGRAM 'LOCATE.BAS'
(, (WRITTEN BY M. ROACH) APPENDIX 5 SEISMIC UNIX SHELL SCRIPT FOR ACOUSTIC FINITE DIFFERENCING
(,
, ~,
('
(" IV r ' List of Figures
Figure Number TitlelDescriptiou Page r' 1.1 Overview map ofstudy area (in the context ofTasmania) 2 1.2 Detailed map ofstudy area 3 2.1 Traverse map and contours ofthe vertical component ofmagnetic intensity 10 for the southern portion ofa magnetic survey conducted in the Derwent Estuary in 1974/1975. ReproducedfromLeaman (1975b). 2.2 Geological cross section from the Bowen Bridge alignment. Modifiedfrom 12 ( , Colhoun and Moon (1984) 2.3 Simplified geological map ofthe Hobart area 13 2.4 A photo ofan exposure ofLower Parmeener Super-Group rocks at Opossum 14 Bay (sparsely fossiliferous marine siltstone) 2.5 A map ofthe distribution ofLower Parmeener Super-Group rocks in the 15 Hobart area , 2.6 A photo ofan exposure ofUpper Parmeener Super-Group rocks at Second 16 Bluff, Bellerive (current bedded sandstone) 2.7 A map ofthe distribution ofUpper Parmeener Super-Group rocks in the 17 Hobart area 2.8 A photo ofTertiary boulder beds exposed on the foreshore at Taroona 18 2.9 A map ofthe distibution ofTertiary sediments and sedimentary rocks in the 19 r-', Hobart area 2.10 A photo ofPleistocene marine sediments exposed at Arm End, South Arm 20 2.11 A map ofthe distribution ofQuaternary sediments in the Hobart area 21 2.12 A photo ofa thermal contact between a dolerite sill and Lower Parmeener 22 Super-Group rocks near Blackmans Bay 2.13 A map ofthe distribution ofJurassic Dolerite in the Hobart area 23 2.14 A photo ofa Tertiary basalt lava flow overlying a tuffaceous deposit exposed 24 at Sandy Bay 2.15 A map ofthe distibution ofCenozoic volcanic rocks (i.e. Tertiary basalts) in 25 the Hobart area 3.J.1 Photo ofthe FMVNubeena, a TAFl research vessel. Taken near Electrona in 30 North West Bay 3.1.2 Photo ofthe FMVPoolta, a TAFl research vessel. Taken near Electrona in 31 ('. North West Bay 3.2 A map ofmagnetic survey lines for 2000 and 2001 used for gridding (i.e. not 32 all ofthe survey lines - some were removed prior to gridding. See also figure 3.9) 3.3 Chart ofvariation in horizontal magnetic intensity measured by the Hobart 34 IPS gradiometerfor 4 days in March 2001, demonstrating the magnetic effects ofa magnetic storm ,'-" 3.5 Chart ofthe G856 base station record for day 2 ofthe magnetic survey, 2001 35 (i.e. 21" ofMarch 2001) 3.6 Profile ofmagnetic intensity on line 5400from the 2001 magnetic survey, 36 showing drop outs 3.7 Profile ofmagnetic intensity on line 5400 after drop out editing 37 3.8 Profiles ofmagnetic intensity (after drop out editing) for allprofiles used for 38 (, gridding 3.9 All survey lines for the 2001 and 2000 magnetic surveys 39 3.10 False colour image oftotal magnetic intensity (TMl) grid 40 3.11 False colour image oftotal magnetic intensity gridfrom regional datasetfor 41 comparison to new data 3.12 Contour map ofTM!, generatedfrom gridded survey data 42 (, 3.13 Schematic image ofthe magnetic response ofa sphere ofsusceptible 43 material, used to illustrate the effect ofboat heading on magnetic measurements 3.14 Heading test data (raw) plotted as afunction oftime. Field and base station 45 records shown
r, V " 3.15 Diurnally corrected heading test data and northings from GPS log as a 46 fUnction oftime, illustrating low resolution ofnon-differentially corrected GPSposition data 3.16 Observed and calculated heading test data. Calculated data used to apply 47 corrections r 3.17 Comparison ofmagnetic data from 2000 and 2001 surveys - selectedpoints 49 from line intersections 4.1.1 Sensitivity testing example chart 1 51 4.1.2 Sensitivity testing example chart 2 52 4.1.3 Sensitivity testing results chart 1 53 4.1.4 Sensitivity testing results chart 2 53 r 4.1.5 Potent modelfor sensitivity testing number 1 54 4.1.6 Potent modelfor sensitivity testing number 2 55 4.1.7 Potent modelfor sensitivity testing number 3 55 4.1.8 Potent modelfor sensitivity testing number 4 56 4.2 Traverse 1 model results 58 4.3 Traverse 2 model results 59 r 4.4 Traverse 3 model results 60 4.5 Traverse 4 model results 61 4.6 Traverse 5 model results 62 4.7 Traverse 6 model results 63 4.8 Traverse 8 model results 64 4.9 Traverse 9 model results 65
(--- 4.10 Traverse 10 model results 66 4.11 Traverse 11 model results 67 4.12 Traverse 12 model results 68 4.13 Traverse 14 model results 69 4.14 Traverse location map (for Potent models) 70 5.1 Photo ofthe boomer seismic source catamaran (on land) 72 5.2 Photo ofthe seismic receiver array (eel) in the water 73 (-, 5.3 Photo ofthe electrostaticprinter used to output the seismicprofiles 74 5.4 Photo ofthe FMVMallanna (the vessel usedfor the seismic survey) 75 5.5 The two boats usedfor the seismic survey at Bridgewater. Two boats were 75 used because ofthe shallow water depths. 5.6 Map ofthe seismic trackpaths, for 2000 and 2001 76 5.7 Calculated magnetic response ofa thin sheet ofbasalt 78 (-" 5.8 Example ofa 'blanket' ofacoustic turbidityfrom Taylor (1992) 80 5.9 Map ofthe distribution ofacoustic turbidity in the Derwent Estuary as 81 definedfrom the seismic reflection proflies 5.10 Example 1 ofa possible phase reversal on one ofthe seismic reflection 84 profiles 5.11 Example 2 ofa possible phase revsersal on a seismic reflection profile 84 5.12 Example ofprobable gas seepage and velocity pulldown (" 86 5.13 Example ofa pockmarkfrom the Derwent Estuary 88 5.14 Example ofapockmarkfrom the literature (Taylor, 1992) 88 5.15 Map ofwater column disturbances definedfrom the seismic reflection 90 profiles 5.16 Schematic ofpossible chemical and biological boundaries and conditions 93 necessaryfor the bacterialproduction ofmethane n 5.17 Map ofthe distribution ofcalcium carbonate in the shallow sediments ofthe 95 Derwent Estuary (provided by Dr. Peter Harris) 5.18 Photo ofthe 'Craic' corer owned by TAP1 and usedfor shallow sediment 97 sampling 5.19 Photo ofborosilicate glass sample bottles usedfor the collection ofwet 98 sediment samples (, 5.20 Map ofthe location ofsediment samples, coded according to measured 99 methane concentration (normalised by sample mass) 5.21 Schematic ofsome seismic energy travelpaths - primary reflections 100 5.22 Schematic ofsome seismic energy travelpaths - simple multiples 100 5.23 Example ofseafloor multiples 101
0 VI (~
5.24 Schematic ofsome seismic energy travelpaths - complex multiples 102 5.25 Example 1 ofmultiple reflections from the Derwent Estuary seismic proflies 103 5.26 Example 2 ofmultiple reflections from the Derwent Estuary seismic profiles 104 5.27 Example ofhow a multiple reflection can inteifere with the interpretation of 105 primary reflections r 6.1 The seismic profile taken alongside the southern side ofthe Tasman Bridge 106 6.2 The input velocity model usedfor acousticfinite difference (AFD) modeling 108 6.3 Highfrequency model output (i.e. a 'pseudo-boomer'record) from the AFD 109 modeling 6.4 Moderate frequency model outputfrom the AFD modeling 110 6.5 Lowfrequency model output from the AFD modeling 111 (' 6.5 Lowfrequency 'pseduo-boomer' record calculated by acousticfinite differencing 7.1 Trackpaths mapfor 'long' seismic lines (i.e. lines 3,5,7,21,22,23,16 and 17) 115 7.2 Trackpaths map 2for 'short' seismic lines (i.e. lines 24 to 34 inclusive) 116 7.3.1 Line 3 raw seismic profile 119 7.3.2 Line 3, interpretation 120 r 7.4.1 Line 23, raw seismic profile 121 7.4.2 Line 23, interpretation 122 7.5.1 Line 22, raw seismic profile 123 7.5.2 Line 22, interpretation 124 7.6.1 Line 5, raw seismic profile 125 7.6.2 Line 5, interpretation 126 7. 7.1 Line 16, raw seismic profile 127 " 7. 7.2 Line 16, interpretation 128 7.8.1 Line 17, raw seismic profile 129 7.8.2 Line 17, interpretation 130 7.9.1 Line 21, raw seismic profile 131 7.9.2 Line 21. interpretation 132 7.10.1 Line 7, raw seismic profile 133 r 7.10.2 Line 7, interpretation 134 7.11.1 Line 24, raw seismic profile 135 7.11.2 Line 24, interpretation 136 7.12.1 Line 25, raw seismic profile 137 7.12.2 Line 25, interpretation 138 7.13.1 Line 26, raw seismic profile 139 (" 7.13.2 Line 26, interpretation 140 7.14.1 Line 27, raw seismic profile 141 7.14.2 Line 27, interpretation 142 7.15.1 Line 28, raw seismic profile 143 7.15.2 Line 28, interpretation 144 7.16.1 Line 29, raw seismic profile 145 7.16.2 Line 29, interpretation 146 r' 7.17.1 Line 30, raw seismic profile 147 7.17.2 Line 30, interpretation 148 7.18.1 Line 31, raw seismic profile 149 7.18.2 Line 31, interpretation 150 7.19.1 Line 32, raw seismic profile 151 7.19.2 Line 32, interpretation 152 ,'. 7.20,1 Line 33, raw seismic profile 153 7,20.2 Line 33, interpretation 154 7.21.1 Line 34, raw seismic profile 155 7.21.2 Line 34, interpretation 156 8.1 False-colour image ofbathymetry, generatedfrom gridded data 160 8.2 Contours ofbathymetry, generatedfrom gridded data 161 ( 9.1 Combined geological interpretation map 165 9.2 Image ofgridded magnetic data, reduced to the pole. Usedfor qualitative 166 interpretation 9,3 Image offirst vertical derivative ofreduced to pole magnetic data. Usedfor 167 qualitative interpretation , VII r· List of Tables
Table Number TitlelDescription Page r 1 List ofsonic velocity rangesfor common geological materialsfound in the 29 Hobart area 2 Results ofheadspace gas analysis 98
(,
r
(.
r·
r
r
(,
(\
(' VIII Chapter 1 Introduction
('
CHAPTERl INTRODUCTION
(', The Derwent River
The Derwent River is one of the largest rivers in Tasmania by any measurement standard. It drains Lake St Clair in the southern central highlands, and meanders over 200km down ( " to Storm Bay south of Hobart where it eventually joins the Tasman Sea.
The modem river deposits most of its suspended sediment load between New Norfolk and (' Bridgewater, and an area of fluvio-deltaic mud flat is developed north of the Bridgewater causeway. South of the causeway is the Estuary proper, and this extends to the entrance of Storm Bay, marked by the significant widening of the water body just south of the Iron
(' Pot. The length of the Estuary between Bridgewater and Storm Bay is nearly 60km, and it is up to 6km wide.
The bulk of this study was conducted in the southern portion of the Estuary, between (" Hobart in the north and Betsey Island in the south, excluding Ralphs Bay to the east and North West Bay to the West. Some seismic data were collected further upstream, as far north as the Bowen Bridge. See the location maps for details.
(" Since the city of Hobart is built on the banks of the Estuary, access is excellent. There are major roads and numerous boat launching and mooring points along both shores of the river. The boats used for this study were launched near Electrona in North West Bay, at the (' Regatta Grounds near the Cenotaph and at the Derwent Sailing Squadron (DSS) in Sandy Bay. Access to the DSS ramp and mooring facilities was arranged by Dr. Michael Roach. The other launching points are freely accessible to the public, via good quality sealed roads.
(,
('
A Geophysical Investigation of the DeIWent Estuary 1 Chapter 1 Introduction
Study Area Overview 300000 400000 500000 600000
~5600000
5600000l0· + + +
! :...
~'~'.~ 5500000-1 +XA. + + ~5500000
t
5400000-1 + \ +' f + +\ ~5400000 ~{.' i'L ".O! &.
4 .-',... & ~
~ ')\,. • 0 0 ~
.~...1~ ~ ..
"
5300000-1' + 1 ~+ \+ 1-5300000 ""\.. + '"
~ ~f 4~~' I + l5200000
I 1-5100000
51000001 I ~ + + + .,~ 1 300000 400000 500000 600000
N 50000 0 50000 100000 Meters I Study Area Coast from 1:250,000 digital coverage A Grid: Australia Map Grid Zone 55 Datum: Australian Geodetic Datum 1966
Figure J. J. General location map ofstudy area.
A Geophysical Investigation of the Derwenr Estuary 2 Chapter 1 Introduction
Study Area 530000
5260000 + 5260000
5255000 + + + 5255000
5250000 + + 5250000
5245000 + + 5245000
5240000 + 5240000
5235000 + + 5235000
5230000 + + 5230000
520000 525000 530000 535000 540000 545000
2500 0 2500 5000 Meters N ~ /\I Major Roads Coast from 1:250,000 digital coverage Land Major Roads from 1:500,000 digital coverage A Grid: Australia Map Grid Zone 55 u::J Study Area Datum: Australian Geodetic Datum 1966
Figure 1.2. Detailed study area map. Note that some seismic profiles were obtained outside ofthe main study area.
A Geophysical Investigation of the Derwent Estuary 3 Chapter I lntroduction
Aims
, The purpose of the study was to investigate some aspects of the geology of the lower Derwent Estuary. This kind of study requires specialised techniques because the rocks are covered by water and mud, ie. direct mapping is impossible. Geophysical techniques enable us to map the hidden geology, provided that certain conditions regarding the (, physical properties of the rocks are satisfied.
Different geophysical techniques exist for the remote detection and/or mapping of different geological features or 'targets'. By limiting the techniques that we use in any given study, we automatically place some restrictions on what we can learn. But we are also limited by the techniques themselves, in terms of the equipment, logistics and geological conditions they require to be used successfully.
As a result, the outcomes that can be reasonably expected are in part dependent on the type oftechniques to be applied, and the techniques to be applied are chosen based on what you are trying to achieve and what you can afford.
Having said that, the aims ofthe project were to determine, where possible: • some aspects of the Cainozoic sedimentation history of the river, and h.alf-graben basin in general • bedrock depth and distribution of bedrock types beneath the Cainozoic sediments • location ofstructures and structural relationships • tectonic and/or geological history of the Estuary (with implications for the greater Hobart area)
There are a number of reasons why this kind of work can be useful. First and foremost it (' can be used as an aid for mapping (ie. to extend the existing maps). Secondly, it can assist in the interpretation of the geological history of the Hobart area. Thirdly, the Derwent is polluted by heavy metals and other contaminants, and knowledge of where sediment has
(' been accumulating in the recent geological past could be useful for pollutant studies. Finally, there is interest in some aspects of the geology of the Derwent Estuary from a geotechnical perspective because the city of Hobart is build around the Estuary. This last
('
A Geophysical Investigation of the DelV(ent Estuary 4 Chapter I Introduction , point is particularly noteworthy, because some of the data collected during this project are already being used as part of a land stability study at Taroona (this is being conducted by
Coffey Geosciences on behalf of Mineral Resources Tasmania, one of the project ,~ sponsors).
Techniques ,.
There are a number oftechniques that are routinely applied in geophysical investigations some of these can be used on land and over water, some are only suitable for land use and some are specifically for use in the marine environment. It is worth briefly considering the (' applicability of the common techniques to this study, and explaining why we chose the techniques that we did. For the purposes of the following discussions I will use the terms 'seafloor' and 'riverbed' interchangeably, and the term 'marine' includes the estuarine
~ environment.
Potential Field (Magnetic and Gravity) Methods
Potential field techniques are efficient (large coverage for low cost), well understood techniques used for both geological mapping and mineral exploration. Measurements of both the magnetic and gravity fields can be done from the ground, sea or air. For these , reasons, we conducted a magnetic survey over the entire lower Derwent Estuary. While it is not particularly difficult to deploy a magnetometer designed for land use from a marine platform, this is not the case for a gravimeter. The specifics of the gravity technique require that acceleration of the deployment platform be removed from the measurements, ,-, and this requires specialised equipment (which we do not have). As a result, we did not conduct a gravity survey. While the magnetic data is informative by itself, the lack of gravity data precludes the use of the traditional coupled interpretation method. This has
{' implications for the interpretation ofthe data, particularly for modeling (see chapter 4).
Electrical/Electromagnetic Methods, including MT and GPR.
( Electromagnetic and electrical methods are used in situations where ground conductivity contrasts can be used for mapping or exploration. They are designed for use over land, where ground conductivities are generally low. The (highly conductive) saline water that
("
A Geophysical Investigation ofthe Derwent Estuary 5 Chapter 1 Introduetion r,
exists in all marine settings has traditionally precluded the use ofelectrical methods, with a few notable exceptions. Magneto-tellurics (MT) and a variation of the MMR method have been applied in marine settings for deep crustal work, but these techniques would not have {' provided sufficient resolution for this study. Time-domain and frequency-domain EM methods for marine use are under development. Some methods (such as GPR) have successfully been applied in areas ofshallowfreshwater cover (Reynolds, 1997).
Seismic Techniques
Seismic methods (reflection and refraction) are very commonly used in land and marine { studies, particularly for petroleum exploration and geotechnical surveys. High resolution seismic reflection profiling is a method used to investigate the structure and sedimentation history of the shallow subsurface in the marine environment, and we applied this technique in the Derwent Estuary. There are a variety of energy sources available for this kind of !' work - the most common are sparkers, boomers and pingers. For this project, we used a 200J boomer source with analogue data recording.
We did not conduct any seismic refraction surveying, for several reasons. Firstly, in order to get data about deep refractors, a large source - receiver offset is required. This requirement is limiting because it can be difficult to keep a sonobuoy string straight and in line with the source when long offsets (and long cables) are involved. This is especially the !' case in the presence of strong currents and leads to uncertainty in positions. Secondly, the expected complex subsurface structure would have demanded a large data volume (sufficient for tomographical inversion) and hence a lengthy, expensive survey. These (' factors suggested that seismic refraction was probably inappropriate for this study. This assessment is in agreement with Learnan (1975), who stated "The refraction technique is not an ideal approach to the problems ofthe River Derwent. .." (page I).
, ' Sonar (Side-Scan, Multibeam and Single-Point)
Similar in principle to seismic reflection profiling but with greatly increased operating frequencies and essentially no sub-bottom penetration are single point bathyrnetric profiling and side-scan (or multibeam) sonar mapping. These techniques are commonly
applied in marine investigations as part of integrated studies, but only provide information ,~
A Geophysieal Investigation ofthe Derwent Estuary 6
Chapter 1 Introduction ,~
about the seafloor (structure ± some compositional information about the surficial sediments). Single point sonar bathymetric data is easily and cheaply obtained with a depth sounder and logging equipment. Side-scan sonar or multibeam sonar require specialised (' equipment. We did not have ready access to side-scan or multibeam equipment, so we collected single point sonar data. This was satisfactory for our purposes.
(
("
(
A Geophysical Investigation ofthe Derwent Estuary 7
Chapter 2 Geology .~
CHAPTER 2 GEOLOGY
( Introduction
The Tasmanian settlements were amongst the earliest European colonisations in Australia,
( and Hobart was settled in the early 1800's (Leaman, 1999). Despite not being known for any mineral or petroleum resources (industrial resources and coal have been sourced from in and around Hobart historically), this relatively long history of European occupation and large population base has meant that the geology of the Hobart area has been extensively (' studied over a long period.
Previous Work
The earliest published references to aspects of the geology of southern Tasmania (and Hobart in particular) can be found in works by Charles Darwin from the mid 1840s, such as Geological Observations on Volcanic Islands (1844) where he makes reference to the basalts exposed around Hobart (Spry, 1955). Darwin gives a more general treatment to the geology of Hobart in his Journal ofResearches into the Natural History and Geology of the Countries visited during the voyage of HMS Beagle round the world, under the command ofCaptain Fitz Roy, RN. (Leaman, 1999).
The first specific guide to the geology ofthe Hobart area was Thomas Harrison's Notes of the Geology of Hobart Town published in the Transactions of the Royal Society of ( Victoria in 1865 (Lewis, 1946). This was followed by R. M. Johnston's Systematic Account ofthe Geology ofTasmania in 1888 (Lewis, 1946).
A number of more specific papers were also published in the late 1800s - these typically related to industrial resources such as coal. The earliest specific work on the Derwent Estuary was also produced around this time - R. M. Johnston's Notes Showing that the Estuary of the Derwent was Occupied by a Fresh-Water Lake During the Tertiary Period (1881).
A Geophysical Investigation ofthe Derwent Estuary 8 Chapter 2 Geology r· Following Johnston's 1888 work, the next - and most comprehensive - was A. N. Lewis' book The Geology of the Hobart District, published in 1946. This gave a thorough treatment to many aspects of the geology, and included a detailed map. (
In the mid to late 1960's a detailed geological mapping program was undertaken by the Department of Mines, and the resulting Hobart 1:50,000 map sheet and accompanying r· explanatory notes Hobart (Leaman, 1976) subsequently became the major reference for the district.
As a more general reference, Burrett and Martin (1989) provides a guide to the geology of ( Tasmania, (including Hobart) in the same vein as Johnston's Systematic Account.
Dr. David Leaman's recent publication Walk Into History in Southern Tasmania (Leaman, 1999) contains updated information on much of the geology of the Hobart area, and is perhaps now the best general guide despite not being a specifically geological text.
Previous Geophysical Surveys
Previous geophysical work in the lower Estuary has been relatively limited. The majority ofthe work has been associated with site investigations for the river crossings, all ofwhich are north of the main study area. Some limited geophysical surveying has been carried out in the lower Estuary, mainly by the Tasmanian Department of Mines in the 1970s.
Seismic refraction surveys were carried out at the sites of the Tasman Bridge (Spry and Carey, 1962) and Bowen Bridge (Leaman, 1975). A seismic refraction survey was also conducted in the area between Kellys Point and the Iron Pot to investigate some aspects of the riverbed structure (Leaman, 1974).
The only previous seismic reflection work in the Estuary was a small 'sparker' survey done for geotechnical purposes - the results were not published (Joe Giedl, pers comm 2001). A major seismic reflection survey just outside of the Derwent Estuary was undertaken by the Amoco Australia Exploration Company in the summer of 1969-1970. This survey obtained approximately 600 miles of24 fold CDP "Aquapulse" data, with the
A Geophysical Investigation of the Derwent Estuary 9 Chapter 2 Geology
('
closest line to the Estuary (TW322) running due east from the southern tip of Adventure Bay on Brony Island through Storm Bay. Arnoco concluded that the thickness of sediment over the "economic basement" was small, and that hydrocarbon potential was unlikely - in ". Storm Bay on line TW322 the thickness of "sediment" (ie. Tertiary and younger) was found to vary from around 300 to 900 feet (ie. 90 to 270m).
(' \,. B ! I RIVER DERWE: MARINE MAGNETIC AREA 1 o soo \lIOO " t...... •d:= ... L -c. ..:::3
GEOPHYSICIST I D.E.LEAMAH Drl1a2/JtlIllUll,: VtBellil.
r'
HOBART
"
.,,., .r~~
'.\,
· '\% N ~'''''>~ '-"'- I' '\ , -t":::::><;:::::::::}" \ Sandy Bay ---,'--" \, , 'l " \ '\. 'I\~\ \\ \
'"', \ '. Y \,j.J~(j1l U~M
,~ Bf/nking Bill )..0 Gmow Paim II / \ '''''I;;:=;() .. .. . Figure 2.1. Traverse map and contours of the vertical component of magnetic intensity for the southern portion ofthe 1974·75 survey, Reproducedfrom Leaman (l975b).
A Geophysical Investigation ofthe Derwent Estuary 10 Chapter 2 Geology
[ Regional, airborne magnetic surveying has been carried out in south-eastern Tasmania by a number of exploration companies, but the only prior high resolution (marine) magnetic surveying in the lower Derwent Estuary was a small scale investigation of some known [, anomalies near Betsey Island in 1973 (Leaman, 1973). In the northern portion of the Estuary, more extensive work has been done. For example, a survey was conducted near Dowsings Point as part of the second river crossing site investigation (Leaman, 1975b), f and a much more detailed study of the northern part of the Estuary (from Sandy Bay to Bridgewater) was undertaken between January 1974 and May 1975 by the Tasmanian Department of Mines. This survey used a McPhar fIuxgate magnetometer mounted on a wooden gimbal to measure the vertical field component. Locations were provided by f' surveying but the surveyed lines were irregularly orientated and the data were not corrected for boat heading. The data were used to produce contour maps, and interpretations of the geology from Bridgewater to the John Garrow light were made by
r Dr. David Leaman (Leaman, 1975b). Some magnetic traverses were also done around the site of the W~est Point casino as part of the site investigation (Stevenson, 1969 in Burrett and Martin, 1989).
There is reasonable ground gravity station coverage around Hobart, but no ship-borne gravity coverage. r The only significant coring within the Estuary was undertaken as part of the river crossing site investigations. Geological cross sections drawn from the core logs are available for both the Tasman Bridge (Trollope et aI., 1966) and Bowen Bridge (Colhoun and Moon, 1984) alignments. Other sediment sampling and coring that has been done by various C' groups (such as TAF1) has been surficiaI (ie. less than one metre core length).
r ,~,
r
A Geophysical Investigation of the Derwent Estuary 11 Chapter 2 Geology
I Western Shore I Arrows mark drill hole I Eastern Shore 30 locations
20
~~ ~ ~~ ~ ~ ~ ~ ~ 10 11
0 Wfjer -10 " •Silt and Cl ay -20 -30 D Dolerite Bedrock -40 D Weathered Dolente -50 D Tnassic Bedrock lOOm -- Scree and Gravel -60 V.E. = 4x Probable D Sandy Silt and Fault Clay (Pleistocene) D Holocene Figure 2.2. Geological cross section from the Bowen Bridge alignment, modified from Colhoun and Moon, 1984.
Geology of the Hobart Area
The following discussion of the geology of Hobart is designed to give a background regarding the types of rocks and sediments that we might expect to be present in the study area. This is based on the assumption that all rock types that might be present beneath the river are present on shore or are known from the coring that has been done. I do not make any claims as to the validity of this assumption.
The exposed geology of the study area (ie. most of the area covered by the 1:50,000 Hobart and Kingborough Geological Atlas sheets - see Leaman, 1976 and Farmer, 1985) consists entirely of rocks of the upper Paleozoic to Cainozoic - there is no exposure of pre Permian rocks in the area. Both Cambrian volcanics and Precambrian rocks are known to underlie the more recent geology (Leaman, 1999). Broadly speaking, there are five groups of rocks exposed in the Hobart region. The oldest of these are Carboniferous and Permian sedimentary rocks of the Lower Parmeener Supergroup. These are disconformably overlain by Triassic sedimentary rocks of the Upper Parmeener Supergroup. Intruding the Permian and Triassic sequences are bodies of Jurassic Dolerite. Attaining a maximum sill thickness in excess of 300m, the dolerite forms many prominent hills and ridges in the modern topography. Unconformably overlying these older rocks are Tertiary volcanic rocks, sedimentary rocks, and unconsolidated sediments of variable association, and the Tertiary sequences are in turn overlain by Quaternary deposits.
A Geophysical Investigation of the Derwent Estuary 12 Chapter 2 Geology
The Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 5265000 5260oo01r~et 5260000
5255000 5255000
5250000 5250000
5245000 5245000
5240000 5240000
5235000 5235000
5230000 5230000
5225000 5225000
520000 525000 530000 535000 540000 545000
Smpliriod G8)logy OJaIOrrNI'Y N T8~ary § Triaalc
5000 o 5000 Meters D P8rmian ~ I Igneous Ftld<. o Tertiary Basalt r=I Jurassic: Oolerite §m'piifed 8(1Jl"1darios N Geologicsl Contad NW.tlerEdgo A Grid: Australian Map Grid Zone 55 I \ 'Fault Datum: Australian Geodetic Datum, 1966
Figure 2.3. A simpLified geoLogicaL map ofthe Hobart area, showing six divisions based on age. This map is derivedfram the 1:250,000 digitaL geoLogy coverage of Tasmania.
A Geophysical Investigation of the Derwent Estuary 13 Chapter 2 Geology
Stratigraphy
Lower Parmeener Supergroup As mentioned, the oldest exposed rocks in the Hobart-Kingborough quadrangle are of the Lower Parmeener Supergroup, of late Carboniferous to Permian age (commonly referred to as Permian). The earliest part of this Supergroup is a sequence of glaciogenic rocks, which are overlain by a thick succession of mainly marine siltstones and mudstones with lesser sandstone, limestone and carbonaceous rocks including some coal. These rocks are variably fossiliferous and form a largely conformable sequence (with disconformities between some Formations). The 'Permian' succession is entirely marine, with the exception of the transitionaVterrestrial Faulkner Group.
Figure 2.4. Permian sedimentary rocks of the Lower Panneener Supergroup exposed on the foreshore at Opossum Bay. The cliffs in this view are approximately five metres high.
Upper Parmeener Supergroup Disconformably overlying the Lower Parmeener Supergroup rocks are those of the Upper Parmeener Supergroup. While Leaman (1976, 1999) assigns the Cygnet Coal Measures to the Lower Parmeener Supergroup, Farmer (1985) asserts that they are in fact part of the Upper Parmeener Supergroup. Both agree that they are latest Permian.
A Geophysical Investigation of the Derwent Estuary 14 Chapter 2 Geology
Permian Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 5265000 52600001[~~ 5260000
5255000 5255000
5250000 5250000
5245000 5245000
5240000 5240000
5235000 5235000
5230000 5230000
5225000 5225000
520000 525000 530000 535000 540000 545000
N Simplified Geology 5000 0 5000 Meters Cl Permian I I Simplified Boundaries N Geological Contact NWaterEdge 1'\1 Fault A Grid: Australian Map Grid Zone 55 Datum: Australian Geodetic Datum, 1966
Figure 2.5. The simplified Lower Parmeener Supergroup (ie. Permian) geology ofHobart.
A Geophysical Investigation of the Derwent Estuary 15 Chapter 2 Geology
The younger, less contentious parts of the Upper Parmeener Supergroup have been split into a number of different associations based on rock type, and range in age from late Permian to late Triassic (commonly referred to as Triassic). These rocks are entirely of terrestrial origin and range from conglomerate to mudstone and coal seams, but the most important rock type is sandstone - either clean quartz, feldspathic or lithic (Farmer, 1985). Often, these Triassic sandstones exhibit current bedding (commonly overturned). Overall, there is a transition up section from clean quartz sandstones to more micaceous and feldspathic rocks, accompanied by a reduction in average grainsize (Leaman, 1976).
Figure 2.6. Triassic sandstones showing typical cross bedding at Second Bluff, Bellerive. Lens cap is 52mm diameter.
A Geophysical Investigation of the Derwent Estuary 16 Chapter 2 Geology
Triassic Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 5265000 52600001I~tl... 5260000
5255000 5255000
5250000 5250000
5245000 5245000
5240000 5240000
5235000 5235000
5230000 5230000
5225000 5225000
520000 525000 530000 535000 540000 545000
N Simplified Geology
5000 o 5000 Meters o Triassic ~ ~ Simplified Boundaries N Geological Contact NWaterEdge 1'\1 Fault A Grid: Australian Map Grid Zone 55 Datum: Australian Geodetic Datum, 1966
Figure 2,7, The distribution ofthe Upper Panneener Supergroup rocks in the Hobart area.
A Geophysical Investigation of the Derwent Estuary 17 Chapter 2 Geology
Tertiary Sediments and Sedimentary Rocks Tertiary deposits (with varying levels of consolidation) unconformably overlie the Parmeener Supergroup rocks at a number of localities in the Hobart-Kingborough quadrangle. There are boulder beds, gravels, silts, sands and clays, and localised occurrences of ferricrete and silcrete (Farmer, 1985). Also of Tertiary age are various deposits associated with basaltic volcanism - notably sub-basalt tuffs and super-basalt gravels (Leaman, 1976).
Figure 2.8. Tertiary boulder beds (matrix supported polymictic boulder conglomerateslbreccias) containing large clasts ofdolerite, exposed on the Taroonaforeshore. Lens cap (centre ofphoto) is 52mm diameter.
A Geophysical Investigation of the Derwent Estuary 18 Chapter 2 Geology
Tertiary (sed imentary) Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 5265000 5260000if~ 5260000
5255000 5255000
5250000 5250000
5245000 5245000
5240000 5240000
5235000 5235000
5230000 5230000
5225000 5225000
520000 525000 530000 535000 540000 545000
N Simplified Geology
5000 o 5000 Meters o Tertiary ~ I Simplified Boundaries N Geological Contact NWaterEdge I '\/ Fault A GrK:l: Australian Map Grid Zone 55 Datum: Australian Geodetic Datum, 1966
Figure 2.9. Distribution of Tertiary sediments around Hobart.
A Geophysical Investigation of the Derwent Estuary 19 Chapter 2 Geology
Quaternary Sediments There are a number of deposits of probable Quaternary age around Hobart. Included among these is the Mary Ann Bay Sandstone, a friable, fossiliferous marine deposit dated by Amino-Acid Racemisation at around 125,000 years old (Murray-Wallace and Goede, 1995). This corresponds to isotopic sub-stage Se, when global sea level was equal to or slightly higher than present. This deposit is presently exposed some 24m above the high water mark at Arm End - implying significant uplift rates. Other Quaternary deposits include talus, scree, silts and beach sands.
Figure 2.10. Fossiliferous marine sands of Late Pleistocene age exposed at Ann End. Lens cap is 52mm diameter.
A Geophysical Investigation of the Derwent Estuary 20 Chapter 2 Geology
Quaternary Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 5265000 5260oo01I~ 5260000
5255000 5255000
5250000 5250000
5245000 5245000
5240000 5240000
5235000 5235000
5230000 5230000
5225000 5225000
520000 525000 530000 535000 540000 545000
N Simplified Geology 5000 0 5000 Meters o Quaternary i I Simplified Boundaries N. Geological Contact N Water Edge I",I Fault A Grid: Australian Map Grid Zone 55 Datum: Australian Geodetic Datum, 1966 r Figure 2. JJ. Distribution ofQuaternary sediments around Hobart.
A Geophysical Investigation of the Derwent Estuary 21 Chapter 2 Geology
Igneous Rocks
Jurassic Dolerite Jurassic dolerite is very common in the Hobart-Kingborough quadrangle. It crystallised from a tholeiitic basalt magma (Leaman, 1976), and varies in grainsize and intrusion form. Differentiation effects can be observed in larger sills and dykes. No Jurassic volcanic rocks are known in this area - the dolerite exposed presently was all emplaced in the (relatively) shallow subsurface. Emplacement was multi-phase, with up to five magma pulses (Leaman, 1999).
Figure 2.12. Intrusive contact between a dolerite sill (lower) and Permian sedimentary rocks (upper) exposed near Blackmans Bay.
A Geophysical Investigation of the Derwent Estuary 22 Chapter 2 Geology
Jurassic Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 5265000 5260000ir~ 5260000 r
5255000 5255000
5250000 5250000
5245000 5245000
5240000 5240000
5235000 5235000
5230000 5230000
5225000 5225000
520000 525000 530000 535000 540000 545000
N Simplified Geology
5000 o 5000 Meters o Jurassic Dolerile ~ I Simplified Boundaries N Geological Contact NWaterEdge 1'\1 Fault A Grd: Australian Map Grid Zone 55 Datum: Australian Geodetic Datum, 1966
Figure 2. J3. Distribution ofJurassic dolerite around Hobart.
A Geophysical Investigation of the Derwent Estuary 23 Chapter 2 Geology
Cretaceous Alkaline Intrusive Rocks These occur only in the far south of the Hobart-Kingborough quadrangle, and are not relevant to this discussion.
Tertiary Volcanic Rocks There is some controversy about the ages of the rocks I have named here as Tertiary (ie. some may be younger), and Leaman (1976) refers to them simply as Cainozoic. I will refer to them as Tertiary (following common usage), but I recognise that this is perhaps not strictly correct. A large number of relatively small volcanic centres are recognised in the Hobart area, and petrographic studies have revealed a range of compositions from typical tholeiitic and alkali olivine basalts to more undersaturated varieties (Leaman, 1976). There are pyroclastic deposits, coherent lavas and minor intrusions associated with the various volcanic centres.
Figure 2.14. Basalt lava flow overlying a tuffaceous deposit, exposed on the foreshore near Blinking Billy Point, Sandy Bay. Lens cap is 52mm diameter. ,
A Geophysical Investigation of the Derwent Estuary 24 Chapter 2 Geology
Tertiary (igneous) Geology of the Hobart Area From 1:250,000 scale digital coverage of Tasmania
5265000 ,5265000 5260oo0i[~6L 5260000
5255000 + 5255000
5250000 + 5250000
5245000 5245000
5240000 5240000
5235000 + 5235000
5230000 + + + 5230000
5225000 + + + 5225000
520000 525000 530000 535000 540000 545000
N Simplified Geology 5000 0 5000 Mete rs o Tertiary Basalt i i Simplified Boundaries N Geological Contact N Water Edge /\1 Fault A GrK:l: Australian Map Grid Zone 55 Datum: Australian Geodetic Datum, 1966
Figure 2.15. Distribution ofTertiary (Cainozoic) basalt around Hobart.
A Geophysical Investigation of the Derwenl Estuary 25 Chapter 2 Geology r- Structure and Structural History
The following notes regarding the structure and structural history of the Hobart area are I- greatly simplified and, of necessity, very brief. Only Jurassic and younger structures are considered here.
I- Work by Berry and Banks (1985) on fault striations of the Parmeener Supergroup rocks concluded that a NW to NNW compressional event formed the fracture pattem associated with dolerite emplacement, and involved strike-slip and possibly reverse movement. This c- event must therefore have commenced prior to the mid-Jurassic, and they suggest that it remained active after the dolerite emplacement. The other major event they interpreted was Cainozoic normal faulting associated with extensional tectonics that led to the development of the Derwent graben, and they suggest that this later event involved " reactivation of many of the older Mesozoic faults. This event occurred in the early to mid Tertiary (Berry and Banks, 1985). Their work did not provide support for an independent fold phase. r- Leaman (1995) does not support the contention of Banks and Berry (1985) that there was regional NW compression prior to or during dolerite emplacement, stating that "Inspection of all faults and contacts, however, does not support this view." (page 153). Rather, he " suggests that the emplacement of the dolerite was part of a rifting (ie. extensional) event that broadly defined the position of the modem Tamar, Coal and Derwent Rivers (Leaman, 1999). He concurs with Berry and Banks regarding the Cretaceous-Tertiary extension. c- Suffice to say that there are at least two generations of faulting and all recent (Tertiary and younger) movement has been normal. There are a variety of fault orientations (dominated by NWINNW and SE/ESE trends), but exposed faults are generally steeply dipping. c', Generally, sedimentary dips are low « 25°) but can be highly variable, particularly in the area ofthe monocline defining the Cascades Fault Zone (Leaman, 1976). co Petrophysical Properties
r
A Geophysieal Investigation of the Derwent Estuary 26 Chapter 2 Geology " Since this study involved the use of magnetic and seismic data, it is useful to discuss 'the relevant petrophysical properties of the materials likely to be present in the study area.
I Magnetic Properties
The following discussion of the magnetic properties of rocks from Hobart is partly based (' on a similar discussion in River Derwent Magnetic Survey, John Garrow Light Bridgewater by Leaman (l975b).
I Parmeener Supergroup (Permian and Triassic) Sedimentary Rocks The Permian glacio-marine sedimentary rocks of the Hobart area are essentially non magnetic, and the same can be said of the Triassic freshwater series. Leaman (1975b) felt that the younger, feldspathic division of the Triassic might have slightly more intense r remanent magnetisation than the older rocks, but not enough to make them particularly magnetic.
Cainozoic Sedimentary Rocks and Sediments r These are regarded by Leaman (l975b) as non-magnetic, and in general this is probably true. However, channels of probable Tertiary age in the Fingal valley that contain significant volumes of dolerite (as boulders and smaller clasts) are apparent in ,. aeromagnetic data and the anomalies appear to be due to the dolerite clasts (Gregg, 2001). Therefore, we could infer that the coarse Tertiary sediments containing large proportions of dolerite around Hobart could also potentially have a significant bulk susceptibility. Sediments with this potential (boulder beds containing large numbers of dolerite clasts) " occur on-shore at Taroona. Note that any magnetic anomalies due to these features would presumably be from induced magnetisation only, since the thermo-remanent components would be randomly orientated (see below for discussion of dolerite). Certainly I agree that
(' the finer sediments such as clays and sands are effectively non-magnetic.
Jurassic Dolerite Leaman (1975b) describes this as "A petrologically and magnetically variable rock with
" intense remanent magnetisation and high volume susceptibility" (pages 2 and 3, after Irving, 1956). In a later work he states that the most magnetically susceptible parts of an intrusion are commonly the granophyric sections - near feeders and/or in the upper section
r
A Geophysical Inve of a sheet, and that the most intense remanent effects typically occur in the fine grained margins of the intrusion (Leaman, 1997), lrving (1956) measured the direction and intensity of remanent magnetisation in 30 samples of dolerite, and found a mean " inclination of -85° (_72° modem value) and mean azimuth of 325° (14° modem value). None of his samples were found to be reversely magnetised. Irving also took measurements 'from 9 irregularly arranged boulders in a Tertiary fault scarp breccia from (' Cartwright Point. He demonstrated that the magnetic orientations within these samples were random, ie. the remanent magnetisation in the boulders had been acquired prior to the breccia formation. Leaman (1975b) gives a range ofsusceptibilities from lxl0-3 to 40xlO-3 SI units, and remanent magnetisation intensities ranging from 0.4 to 7 A.m- l directed at r 85°/325° (after Irving) for the Jurassic dolerites. Tertiary Basalts r Leaman (l975b) states that the "...range of properties of the basalt are considered to, be greater than for the dolerite." (page 3), implying a large range of values from highly magnetic to effectively non-magnetic. However, he goes on to state that the variation is dependent on the degree of weathering. The basalts exposed around Hobart are all deeply r weathered, and the weathered basalts tend to have lower susceptibilities than the dolerite. To this, we can add that some of the basalts are reversely magnetised - ie. they have a thermo-remanent component that is (broadly speaking) opposed to the modern field r orientation. This means that they can appear as anomalous lows in magnetic survey data, However, there are no published data regarding the intensity or orientation of the thermo remanent component of the basalts. Since the geomagnetic field changed orientation numerous times in the Tertiary, there are likely to be basalts with their remanent r components aligned with the modem field, ie. they are not all reversed. Acoustic Properties (ie. sonic velocities) r~, The following table of sonic velocity ranges is taken from appendix I of Leaman (1975a), The River Derwent: Elwick Bay to Macquarie Point. Geology: Facts, deductions and problems and is relevant to the discussion of seismic data. r r A Geophysical Investigation of the Derwenl Estuary 28 Chapler2 Geology " Material Seismic Velocity (m/s) Water 1500 Cl River Silt (suite I) -1400-1500 Sand, clay (suite 2) 1600-1800 Weathered Basalt 1700-2500 Basalt (fractured-massive) 2200-6000 (") I Weathered Dolerite 1700-2500 Dolerite (fractured-massive) 2200-6000 Sandstone (weathered-massive) 2000-4500 () Siltstone (weathered-massive) 2000-5000 Table 1. Sonic velocity properties. Leaman (1975a) also states that prior surveys had suggested a velocity of less than 1500 (') m/s for some river silts. (") () (") n o (") A Geophysical Investigation of the DelWent Estuary 29 Chapter 3 Magnetic Survey CHAPTER 3 MAGNETIC SURVEY Field Work 2001 th th st The survey was conducted over three days (the 19 , 20 and 21 ) in March 2001. East west lines at 300m line spacing were surveyed, with navigation by differentially corrected GPS. North-south tie lines were also surveyed. Diurnal corrections were applied using a base station record collected at the Tasmanian Aquaculture and Fisheries Institute (TAA) labs at Taroona with the EG&G Geometrics G856X magnetometer. The roving magnetometer was a GEM Systems GSM-19 Overhauser magnetometer. Both of these instruments measure total magnetic intensity (TMI). The GSM-19 was set to fast sampling l auto mode with a one-second sample interval. At boat speeds up to around 4Okm.hr- , this equates to a maximum spatial sampling interval (along line) of around Ilm. The vessel used for the data collection was the FMV Nubeena, an aluminium monohull approximately 16 feet long. Figure 3.1. The FMV Nubeena at North West Bay. The GSM-19 sensor would nonnally be mounted on a pole attached to a backpack and used on land. For this survey we mounted the sensor on a long aluminium pole, which was A Geophysical Investigation of the Derwent Estuary 30 Chapter 3 Magnetic Survey then lashed to the vessel. When mounted in this fashion, the magnetometer sensor extended several metres from the front of the boat. The sensor itself and the cable connections were wrapped in plastic and sealed so they would be watertight, to protect against spray, rain and possible dunkings. 2000 Magnetic data were collected in 2000 as part of a trial survey - the instruments were the same as for 2001, but a slightly smaller vessel was used. The s'urvey lines were not regular as in 2001, see figure 3.2 (next page, also figure 3.8 page 38). Bathymetric data were not collected in 2000, but were collected in 200 1 (see chapter 8). The survey was run from the 25 th of August to the 28th of August. Figure 3.1.2. The FMV Poolta (another TAFI research vessel), with the GSM-19 magnetometer mounted in place and sealed for water-tightness. Note that the pole is not fully secured in this photo. An identical mounting method was used on the FMV Nubeena. A Geophysical Investigation of the Derwent Estuary 31 "' 525000 530000 535000 540000 ,-.. 5250000 + + + 5245000 + + + (' " 52400[)[) + + + (~ 5235000 + + + r 5230000 + + + + ~.--, 0.0 2.0 4.0 "6.0 8.0 10.DIem 1-.-. i (' School 01 Earth Sciences Localion 01 All (' Grid: AMG Zone 55 Uagnetic Trllvcrses Dalum: AGO 1966 °li~ . .. DrBwn By: D.Gibbons 06-11-2001 Figure 3.2. Location map ofall magnetic sun'ey lines for both the 2000 and 2001 surveys. The 2001 lines are east-west and north-south (ties) and the 2000 lines are north-east and north-west. 32 Chapter 3 Magnetic Survcy r-. Space Weather The purpose of collecting magnetic data from a geological perspective is to gain an insight i into the sub-surface distribution of magnetic susceptibility. This can then be combined with geological knowledge to produce an interpretation of rock type distribution and geological structure. This means that we, as geologists, are generally only interested in o magnetic fields associated with the geology. Naturally occurring fluctuations in the earth's magnetic field due to non-geological sources need to be removed from our data in order to ensure we are only considering the magnetic fields associated with the geology (and the main field of course). Such fluctuations can be generated by thunderstorm activity and by o ionospheric currents resulting from the interaction of the solar wind with the ionosphere. Occasionally, solar activity increases and the solar wind flux increases accordingly. A .~ peak of solar activity occurs approximately once every 11 years. This heightened solar activity can result in abnormally strong ionospheric currents - producing auroras and interfering with geophysical techniques. Techniques that can be affected by these currents include some electromagnetic methods and magnetics. The magnetic expression of r- abnormal ionospheric currents is a component of magnetic intensity that varies rapidly, both spatially and temporally. It may have large amplitudes (ie. greater than the geological component). Such variations cannot typically be removed from survey data by a diurnal r. record due to their rapid spatial variation. A period of heightened solar activity occurred during the magnetic survey in 2001. Fortunately, the most extreme fluctuations occurred during the night rather than during the i day when the data were being collected. IPS Radio and Space Services (a division of the Federal Government Department of Industry Science and Resources) maintains a magnetic gradiometer in Hobart that '" measures the horizontal and vertical field components (plus the declination) every second, on a continuous basis. ,~ r A Geophysical Investigation of the Derwent Estuary 33 Chapter 3 Magnetic Survey Raw Gradlometer Data (courtesy of IPS Redlo eo. Space services) lOO~------ 50 I=" S -SO i i-lOO § "C ! ·150 f .ll -200 3 !I :z: -250 .;lOO _r 1 o 24 48 72 96 120 Tlme (hou.. I"'1t 11"," 1~1) Figure 3.3. The horizontal magnetic field component as measured by the IPS gradiometer (see text) for aroundfour days from the 18th to the 22nd ofMarch 2001. Boxes indicated approximate survey times. The horizontal field component as measured by the IPS gradiometer is shown in figure 3.3 as a function of time. Note the relatively low frequency, low amplitude variation towards the left of the chart (0 to 48 hours), and compare this to the high frequency, high amplitude variation later in the week (commencing at around 64 hours) associated with the storm activity. Although the amplitude of variation is fairly high it is relatively systematic, and therefore could be removed using a base station record. The base station record for day 3 shows nearly 500nT variation in total magnetic intensity (TMI) - this is nearly an order of magnitude more than "normal". A Geophysical Investigation of the Derwent Estuary 34 Chapter 3 Magnetic Survey Raw Base Station Data 621300 ,----_-----~-- 621200 621100 621000 1=' .E. ! 6201lO0 ,-. 620700 62~01 1 o 50 100 150 200 250 300 350 400 450 500 nme(mlnut.. put 8:20:15 AM 210t Morcll2001) Figure 3.5. The G856 (base station) record for the 21st of March 2001, as measured at the TAFllabs at Taroona. Figure 3.5 shows the base station (G856) record for the 21 st of March. Note the large but systematic variation in TMI. If the survey had been conducted several hours later or earlier on the 21 S\ the data might have been terminally corrupted by the magnetic storm. At one point it looked as if the survey might need to be postponed, but fortunately this was not necessary. Clearly, this survey was an example of a near miss and demonstrates the prudence of checking the space weather forecasts! Data Processing Typically, geophysical data requires some processing before they can be used constructively. For magnetic data, this might include diurnal corrections and various forms of data editing such as tie line levelling. The data for both the 2000 and 2001 surveys was diurnally corrected, but more processing was required before it could be used. One of the problems associated with the 2001 magnetic survey data was the relatively th large numbers of "drop-outs" experienced when the swell was strongest (on the 20 ). As mentioned, the sensor was mounted on a rigid pole. Although the pole was lashed down as tightly as possible, during heavy swell conditions the pole would occasionally lift slightly and then slap on the deck. The jarring induced by this slapping caused the signal to drop A Geophysicallnvesligation of the Derwenl Estuary 35 Chapter 3 Magnetic Survey r out. After day 2, the pole was padded to reduce the number of drop-outs and the data quality improved noticeably. .-. Fortunately, drop-outs are relatively easy to remove from the data because they are always negative, and have relatively high frequencies compared to most geological features (often the drop-out is only a single sample). The aeromagnetic processing program ChrisDBF n. was used to remove drop-outs from the raw survey data. This program enables the use of linear or cubic-spline interpolation for profile editing. Rarely, it was difficult to determine whether a particular negative excursion on a profile was geological or a drop-out, so some subjective data editing occurred. Examples of raw and edited profiles are shown below. r Some drop outs were also present in the 2000 data, and these profiles were also edited. "ij.1!1i¥" ,jFiI, +,,,ixl r r-~"-"''''"'"' ,-~"'''.'' "'~c,,""'" I . i'I. ~,.,cr-o,-,,,,r r· 0. Figure 3.6. Line 5400, prior to drop-out editing. Note the high frequency, negative excursions superimposed on the otherwise lowfrequency profile. These negative excursions are drop-outs. re t). n (" A Geophysical Investigation ofthe Derwent Estuary 36 Chapler 3 Magnetie Survey "~ "PU!i¥"I.iWbU -;:!"nx! c r'. ,. Figure 3.7. Line 5400, after drop-out editing. Note the removal ofthe high frequency negative excursions. The final gridded dataset is shown as a series ofmaps on the following pages. The regional ,. dataset is also shown (for comparison). r. , ~, ,~ . r .~ A Geophysical Investigation ofthe Derwent Estuary 37 ~ 525000 530000 535000 540000 ,~ 5250000 + + r 5245000 + + ,~ (' 5240000 + (J~ r 5235000 + + , r 5230000 + + + + 0.0 2.0 '.0 6.0 0.0 10.0km /"'-..... r' School of Earth Sciences Profiles 01 ,- Magnetic Traverses Grid: AMG Zone 55 Dalum: AGD66 Qili Drawn By: D. G. 06-11-2001 ~, Figure 3.8. Profiles oftotal magnetic intensity for selected traverses from the 2000 and 2001 surveys. Note the extremely high amplitude anomaly near Betsey Island. 38 525000 530000 535000 540000 ',----, 5250000 + + + , 5245000 + + + ('"""'.. r 5240000 + + + (J~ r 5235000 + + + o " N 5230000 + + ~, 0.0 2,0 4,0 6,0 8.0 10.0km ~·_·I r. School of Earth Sciences Ijj Location of Gridded ( ...... Magnetic Tnlverses Grid: AMG Zone 55 Datum: AGD66 Bioi :Dt:: Drawn By: D. G. 06-11·2001 " Figure 3.9. Location ofgridded magnetic traverses - i.e. the traverses from the 2000 and 2001 surveys used to create the grid oftoto/magnetic intensity shown infigure 3.10 (page 40). 39 525000 530000 535000 5~0000 5250000 ~ + ~ + + ~""~l. [ tJ 52~5000 U :i ~~.+' ? + \ + 52~0000 + + + 5235000 5230000 + + + 0.0 2.0 ~.O 6.0 8.0 10.0km ~--~ 6955~.9 68585.2 67550.8 66516.~ School of Earth Sciences 65~82.1 Total Magnetic Intensity 6~~~7.7 ~ Histogram Equalised 63~13.~ West-East Sun Angle 62379.0 TMAMG55. AGD66 ~~613~~.6 QB1I~ , Drawn By: D. G. ~ 06-11-2001 Figure 3.10. Image ofTMI from gridded data (2000 and 2001). Grid cell size is 50m; units are nanoTelsas (nT). The grid was interpolated using a spline algorithm and 650m scan distance in ChrisDBF. 40 525000 530000 535000 540000 5250000 5245000 5240000 5235000 5230000 0.0I·.·.·' 2.0 4.0 6.0 8.0 10.0km 14275.7 13052.9 11748.6 10444.3 School of Earth Sciences 9139.9 Regiona'TMI 7835.6 It] Histogram Equalised 6531.3 NW Sun Angle 5226.9 TMAMG55. AGD66 0 3922.6 ·RJI· Drawn By: D. G. 06-11-2001 Figure 3.11. Image ofTMI from regional grid. Units are nanoTelsas (nT), base has been subtracted. Note the lack ofresolution compared to the new data (figure 3.10) 41 525000 530000 535000 540000 -;; 5250000 ~ + \.. .. ·r t{.;-;1'~ 'i\ \ + I ((+l \J f clt1t LJ ~ ( + 5245000 -+ 0 + > -- .Ill + 5240000 + + ~ + \.. ~ 5235000 + ~+ . / ~J~l' ~II ..~~ ,-~ o JU-lllJ "1'<1 ~ 5230000 + + + 0.0 2.0 4.0 6.0 8.0 10.0km ~--·I r.... III Contour Specifications Interval From To Pen School of Earth Sciences 50 61300 61500 50 61550 61700 Contours of Total 50 61750 61900 50 61950 62100 Magnetic Intensity 50 62150 62300 100 62400 62800 200 63000 64000 Grid: AMG Zone 55 500 64500 66000 1000 66000 70000 Datum: AGD66 QB~ Drawn By: D. G. 06-11 -2001 I Ib:::: r Figure 3. 12. Contours oftotal magnetic intensity generated from the gridded magnetic survey data (see figure 3. 10 for a false colour image ofthe grid). Units are nanoTelsas. 42 Chapter 3 Magnetic Survey Heading Test One of the difficulties associated with deploying a magnetometer from a marine platform is the ferrous material (ie. iron) associated with the platform. Unless the scale of operation is very small (ie. a rowboat or canoe) some iron will almost certainly be present, particularly in the boat engines. Although the sensor is mounted away from the boat, a component of magnetic intensity measured at the sensor is due to the vessel. Furthermore, the measured component varies according to the orientation of the boat. This is because the main field of the earth (the inducing field) is a vector field, and the induced magnetic field associated with a susceptible body will be aligned with the main field even if the body is rotated. So as the boat changes direction, the measured component of magnetic intensity due to the boat changes. Figure 3.13. Image of TMI calculated for a sphere of magnetically susceptible material (see text). Colours indicate relative intensity (red = high values, blue = low values). North is grid north. The anomaly peak is not centred on the causative body because of the declination (14 0) and inclination (-72 0) of the inducing field Figure 3.13 shows the magnetic response of a sphere of susceptible material of one metre diameter, and 0.1 SI units susceptibility (the black dot). The image has extents of around 30 by 30 metres and the inducing field has a magnitude of 62000nT. The induced magnetic response associated with the sphere is mildly asymmetric, and the maximum is displaced A Geophysical Investigation of the Derwent Estuary 43 Chapter 3 Magnetic Survey relative to the centre of the sphere (these effects would be more significant at lower magnetic latitudes). If we imagine that this sphere represents the motors of the FMV Nubeena, and the black circle represents all possible sensor positions (relative to the C' motors), then we can see how the sensor will measure different field strengths associated with the motors depending on their relative positions. I' We can quantify this in terms of field strength as a function of angle - ie. the angle of the long axis of the boat relative to a datum angle, such as magnetic or grid north. The long axis of the boat is used to calculate the angle, because the bulk of ferrous materials (ie. the motors) and the sensor are mounted in line with this long axis. Furthermore, the boat travel r direction (expressed as an angle relative to north) is the same as the angle of the long axis of the boat, assuming the boat is moving forward along line, which it was at all times during data collection. r Since the interest lies only in the magnetic response of the rocks (and not of the boat, or magnetic storms!), it is useful to correct for effects such as the one described above. This requires a heading test, which is a means of quantifying the effects of the orientation of the deployment platform. It is done by rapidly sampling the magnetic field, while slowly rotating the deployment platform relative to the sensor, which should remain in a fixed position throughout the test. In this case, the test was performed by driving the boat into i- shallow (knee-deep water) and raising the motors. The sensor was held firmly in place while the boat was slowly pushed around the it. Several rotations were performed and a GPS record was obtained simultaneously. Note that the sensor remained attached to the mounting pole throughout this exercise. I' C' C' A Geophysical Investigation of the Derwent Estuary 44 Chapter 3 Magnetic Survey Heading Test Raw Data I-Aeld -8.... 1 62055 62050 62045 6200) ~ ,, ! 62D35 62030 .f"\\= / l/!t.C- l '" -1 62025 62020I, I r- 104550 104600 10-«l5O 104700 104750 104a00 104650 104900 104950 105000 Tme (O.condO) Figure 3.14. Raw field (blue) and base station (red) recordsjor the heading test. Figure 3.14 shows the raw heading test data, plotted as a function of time. The data was diurnally corr~cted using the base station record prior to any further processing. There are nearly three full boat rotations shown here - note the periodicity of change in orientation with time (in this case, a proxy for angle). The variation in intensity with boat orientation covers a range of around 22nT. It is useful to conduct the test in an area with low magnetic gradients, so any inadvertent lateral movement in the sensor does not significantly impact on the results. Our heading test was performed in the northern end of North-West Bay, where the gradients are relatively low. Unfortunately the test was not performed with the FMV Nubeena, but with the FMV Poolta instead. The Nubeena had been decommissioned by the time we performed the heading test. This will have some influence on the applicability of the heading test to the data collected with the Nubeena, but both vessels have motors of similar size and design so the difference should not be substantial. Both vessels are constructed of aluminium and carry little ferrous material apart from that in the motors. ,~ r A Geophysical Investigation of [he Derwent Estuary 45 ChapLer 3 Magnetic Survey TMI and Northing l..... rtO<1tl!rlg ...... DlII 5237614 20 5237612 - 15 r 5.237610 10 5237608 5 [5.237606 ;:: ... ~ .5 0 i ~ 5237604 ~ r -5 5237602 I -10 5237600 5237598 -15 5237596 -20 " 104'150 104511 104577 104643 104707 104774 104840 104907 104972 Tlm. (....ond.) Figure 3.15. Plot of the diurnally corrected TMI against time (a proxy for angle), with northing (from the GPS log) for comparison. Changes in peak time compared to chart 3 result from a time shift required to synchronise the GPS and magnetometer records. Figure 3.15 shows the diurnally corrected heading test record and northing from the GPS log. The GPS positions could not be differentially corrected because we couldn't receive the differential beacon signal, so the resolution is poor. This lack of resolution means that direct calculation of the boat orientation from the GPS record would not provide a satisfactory result. Further, this method would mean that estimates regarding the absolute sensor position and sensor-GPS antenna separation would be required - these could not be made (without significant errors) from this GPS log. Fortunately there is another way to calculate the angles without reference to the GPS log. Theory dictates that the maximum response due to the boat motors would be experienced with the sensor due north (ie. magnetic north) of the motors. This is also demonstrated from the calculated response due to the sphere shown earlier - the peak response of the anomaly is offset to the north by the inclination, and to the east (away from grid north) by the declination of the main field. Given this, it is possible to assign angles to the measured response based purely on the amplitude of the magnetic field. The method used involved calculating a sinusoidal function of angle, and fitting it to the measured response over one rotation interval. It was not possible to fit the function to multiple rotations simultaneously because they were performed at different speeds, meaning phase differences became significant. A Geophysical Investigation of the Derwent Estuary 46 Chapter 3 Magnetic Survey Calculated and Measured Responses 1-Calculldsd -Measured I 15 TI------~~--~=~ 51 / ~~ i=' i.,!; 0 I'' I,}• rU \' , ... ·5 I ~.",-- / ,I l7.l..l: I 1 /" ·15 oI" , , . , 80 120 180 240 300 360 420 4110 540 600 860 720 790 Angle (181011II010 magnlltlc north) Figure 3.16. The measured and calculated changes in TMI as functions ofangle. Note that although this plot 0 shows a peak value at around 460 ~ I shifted the calculated peak value to 0 before applying the corrections. Figure 3.16 shows the results of fitting a calculated curve to the observed data. The fit was calculated over the interval 280 degrees to 640 degrees (one full cycle), and this region was used to produce the look up table of correction as a function of angle. Note the phase difference between the calculated and 0 bserved curves for the other cycles due to the difference in rotation speed. Although the angles were only calculated from the one cycle, the amplitude of the sinusoidal function was chosen by assessing all three rotations. The advantage of using the sinusoid is a regular and rapid sample interval, and the elimination of high frequency noise. The variation between the measured and calculated responses is never more than 2nT (-10%) over the fitted interval and typically much less. The measured response shown in figure 3.16 is in a diurnally corrected but raw state - if it had been used to apply the corrections (rather than the calculated response), it would have required filtering to remove some of the high frequency variations. The filtering process would have produced an output which would more closely approximate the calculated ~ response than the raw observed data does (in terms of RMS error). Once the look up table of correction as a function of azimuth was calculated, a Fortran 77 ,- program (heading) was written to apply the heading test corrections. This program uses publicly available subroutines to interpolate (using cubic splines) a correction for any arbitrary angle based on the look up table values. Angles relative to grid north were A Geophysical Investigation of the Derwent Esruary 47 Chapter 3 Magnetic Survey r, calculated using the arctangent of the change in northing,' change in easting ratio, and then corrected for declination, The source code for this program (headJ) is at appendix 2, r Once the diurnally corrected, drop-out edited data was heading test corrected, it was re gridded. The same heading test corrections were also applied to the magnetic data collected in 2000, which used a different vessel again (but still aluminium, with twin two r, stroke outboard motors). Although the variation in magnetic intensity due to boat direction changes is small relative to the larger geological variations - tens of nanoteslas compared to thousands of nanoteslas r - the heading corrections did result in a noticeable reduction in data striping in regions of low magnetic gradients, and improved the meshing of the two surveys (note these had different survey line orientations - east/west for 2001, and dominantly north-east/solith .r west for 2000). On that note, it is worth mentioning the variation between the two surveys. The magnetic data collected in 2001 was of generally good quality, with regular line spacing and (' orientation. The data collected in 2000 was of considerably lesser quality, with irregular line spacing and orientation. The two surveys do not mesh together well. To demonstrate this, I have selected a line from the 2000 survey that crosses several lines from the 200 I (; survey. I checked the measured TMI at the intersection points after heading test corrections and drop out editing, and plotted the difference in TMI between the 200 I survey and 2000 survey for each intersection point. The results are shown below. (', () () () A Geophysical Investigation of the Derwent Estuary 48 Chapter 3 Magnetic Survey Comparison of 200112000 Magnetic Surveys 300 200 _100 S :i ~ ! ~ ·100 =::I ~ o -133 :;: -200 -300 .. -400 o 3 Comparison Number Figure 3.17. Not only is the difference in TMI inconsistent across the different lines, it also covers a very large range: from -391nT to 221nT, or a total of 6J2nT! I can only attribute this massive discrepancy to poor data quality for the 2000 survey. The 2000 survey does not grid well in an unmodified form (ie. it requires significant levelling), and even after removal of some of the 2000 survey lines the combined grid of both surveys is low quality in the northern section (where there are no 2001 data). For this reason, the image and contours generated from the grid appear "noisy" in the north. The interpretation and models from the northern portion of the study area are also less reliable as a result, even though the heading test corrections did improve things somewhat. The gridded images (and contours generated from the grid) were generated usmg ChrisDBF. The grid has a 50m mesh, and was made using a spline interpolation algorithm with a scan distance of 650m. A Geophysical Investigation of the Derwent Estuary 49 Chapter 4 Magnetic Modeling r CHAPTER 4 MAGNETIC MODELING I Introduction One of the great advantages of potential field data (magnetics and gravity) is the availability of modeling programs that enable real time calculation of complex model (' geometries. One of the great downfalls is that it is virtually impossible to find a unique model solution for any given real-world potential field measurement. It is vitally important to constrain models as tightly as possible by utilising existing geological data such as I outcrop locations, and by using realistic petrophysical properties measured either on the rocks whose geophysical response is being modelled, or by equivalent rocks from nearby areas. If possible, combined interpretations of magnetics and gravity are preferred to interpreting one in the absence ofthe other. C' In this case, there are no hardrock core logs south of the Tasman Bridge and no gravity data is available, so the modeling performed on the magnetic data is poorly constrained. I (- have used realistic magnetic susceptibilities and remanent magnetisation intensities, but the magnetic properties of the Tasmanian dolerites and basalts overlap significantly (see chapter 2). ,~ Potential field modeling should, ideally, be constrained according to the five criteria described by Leaman (1994). He states that acceptable solutions should: 1. Honour geologic or other control I 2. Not contain discontinuities in geometry or properties that are not justifiable 3. Use rock properties (ie. magnetic susceptibility, density) that are within known ranges or that can be inferred from the data 4. Be based on several interlocking profiles across the data set n 5. Have a consistent (or correct) base level for the entire data set I have attempted to meet these criteria where possible, but the particulars ofmy study have meant that criterion 1 has little meaning, and I have not met criterion 4 due to the (' difficulties associated in modeling along strike using 2D models (where the 'infinite extent perpendicular to section' approximation does not apply). r A Geophysical Investigation ofthe Derwent Estuary 50 Chapter 4 Magnetic Modeling Sensitivity Testing Sensitivity testing is a means of assessing the suitability of various model solutions during forward modeling of potential field data. A sensitivity test can be performed by varying a model parameter and assessing the change in fit between the observed and calculated data. The sensitivity can then be assessed by plotting the (systematically varying) model parameter against a measure of the model fit (such as RMS error). Sill: 200m thick, 350m depth to top. 70, I 001 A sol"- /" 1 ~ I ...... >/ 1 ~~. - I ~ £ g ~ • ~301...... a:: -._ 20 I ...... " I 10 I ...... =-'1 o I 1 o 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05 Magnetic Susceptibility (SI) Figure 4.1.1. In this example, the parameter being varied is magnetic susceptibility. This is for a small section ofthe 2001 data, modelled as a sill. In this case, the sill is modelled as being 200m thick, and with a depth-ta-top of350m The RMS error reaches a minimum at a susceptibility ofaround 0.024 SI units - this is the susceptibility that produces the best model fit for the specified geometry. The model sensitivity is approximated by the slope of the sensitivity function adjacent to the minimum. If the function has a shallow slope adjacent to the minimum, this means that a small change in the modelled parameter has only a small effect on model fit, and therefore the model is not very sensitive (to changes in that parameter). On the other hand, if the function has a steep slope adjacent to the minimum, then a small change in the modelled parameter will have a large effect on the model fit, and therefore the model is highly sensitive (Roach, 1994). A Geophysical Investigation of the Derwent Estuary 51 Chapter 4 Magnetic Modeling Hypothetical SIII8 I..... LowK_HlghKI 90 eo -- . 70 .-.--- ~ 60 ----- / loo / ..g ~40 '\ / a: . 30 ... " '\ / . ~ , ...' ,_,-" - - , , 20 \/ 10 I J __ • ••• _11 _ • _____ ...... _____ '-../ o .. - o 50 100 150 200 250 300 350 400 450 500 Depth 10 lop (m) Figure 4./.2. /n this example, the sensitivity ofdepth to top for two hypothetical sills has been plotted. The "High K" sensitivity function has steep slopes adjacent to the minimum, and therefore the model is sensitive to small changes to sill depth. The "Low K" sensitivity function has shallow slopes adjacent to the minimum, so the model is not highly sensitive to sill depth. In general terms, a highly sensitive model solution is preferred. This is because highly sensitive solutions permit only a small range of feasible properties and geometries compared to insensitive solutions. Of course, for complex models or where geological constraint is poor, the sensitivity testing method described above is inadequate. This is because of the multi-dimensional nature of the sensitivity functions. For example, model fit will typically be influenced by body geometry (shape and orientation), body size and susceptibility. For a simple body (eg. a dyke), this could be broken down to: depth to top, width, dip, strike extent and susceptibility (i.e. five parameters). Testing of each parameter individually, while keeping the others constant, would be unfeasibly time-consuming. To illustrate the sensitivity of my models (actually, the lack thereof), I have conducted limited sensitivity testing on a geologically simple section. I deliberately chose a transect that could be modelled effectively with one* simple body. This section is from the southern portion of the field area, and can be modelled reasonably well with a single sill of dolerite (*plus one other body almost off section that was not varied at all). I calculated A Geophysical Investigation of the Derwent Estuary 52 Chapter 4 Magnetic Modeling sensitivity functions for various sill thicknesses, depths and magnetic susceptibilities. The results are shown below. Susceptibility =0.018 (SI) I...... 200 !hick --300m !hIck lOOm Ihlck -'4OOn Ihlck I 120 100 80 l Ul 60 ::I a: 40 20 0 0 100 200 ~ 400 500 Depth ID lop (rn) Figure 4.1.3. Sensitivity test results A - constant magnetic susceptibility. See below for discussion. Depth to top. 350m l---l00m Thick __200m Thick 300m Thick __·400m ThiCk I 8O,------~- ... 70 1._ _- " __ I 60 I " " ...." 5O.p...~...... / >C I l __ rt' a:~401 ,<...... oc:::::::: ,X" • ...... 1 30 I 'at '...... """-::7' ...L I 20 1 -'<:::"c,' ...... >'" I 10 1 ,,,...... c= ~ < r ...... :;R -...... -. .::+:="" 4 o 1 , o 0.01 0.02 0.03 0.04 0.05 Magnellc Susceptibility (SI) Figure 4.1.4. Sensitivity test results B - constant depth to top. See below for discussion. The sensitivity function curves shown above indicate two things. Firstly, since the slopes of individual curves are all relatively shallow adjacent to their respective minimum, there is a range of feasible values for each parameter for each model. Secondly, the minimum A Geophysical Investigation of the Derwent Estuary 53 Chapter 4 Magnetic Modeling for adjacent curves have similar values - so the model fit for a 200m thick sill at 350m depth and with a susceptibility of around 0.0225 SI is equivalent to that for a 400m thick sill at 350m depth with a susceptibility of around 0.013 SI. Equally, both of these solutions are equivalent to that for a 300m thick sill at 400m depth with a susceptibility of 0.018 SI and so on. Furthermore, all of these solutions are geologically feasible and none can be disregarded based purely on geological arguments. It is also perhaps worth re-iterating here that this is a particularly simple model (effectively only one magnetic source), and that in terms of body geometry, only the thickness was varied in this testing. When more complicated shapes are modelled, the range of available solutions increases. We can generalise by saying that: since there is effectively no geological control for the lower Derwent Estuary, and there is a wide range of magnetic properties for the materials that are likely to be present, the modeling results should be regarded with caution. Another way of considering the equivalence of varying model solutions is by constructing models with equivalent model fits but varying source geometries and/or properties (Roach, 1994). The following four models are all constructed with uniform model susceptibilities only the geometries change. M?tMl"Pt'ti'!!!dM!·Jeeolt1t,MSltHlitfflkM ;mE! DIia'lliiIlIltl f IW.t\;latt~111'l,."Q.LI '000" ( -~------ TMI(,.l) P(m) """ ""'" ~- .2 t 1 ·200 I -: .'00 I '-.. E 2(m) A;r. tJ7l1d'll BOO For~J~FI Figure 4.1.5. Example #1 ofalternative model geometries producing equivalent observed responses. (~ A Geophysical Investigation of the Derwent Estuary 54 Chapter 4 Magnetic Modeling )( RfBle. ~.!POis.'I'M._ t!Il> ~.!!J ·DI~lglialf~Jotl~T¥lST~r t '000» I ,,' .l'~ ------.....-~~_-~ rWI(IIT) --- .. .1ClOO.O p(m) 2000 ""'" _.-+ ..2 3 - •• ·""1 .,r~'" ~--- ...-....<. ...:...... :..: -' 000 . • .. ! I l>:m) : I J \ '~.;:';", ~ 1 h I _.. I g.,,,g, I ~31."'7IjJO""""".l!i"S232031J Figure 4.1.6. EXample #2 ofalternative model geometries producing equivalent observed responses )( I:JBIe _- _.~ r'" I'M _ I!*> ..J4JiJ .QJ~lgl al'f!rt;~JotIG1ci:J~&Jr r 1000.0 l/ 1/ '~ ...~... 00 ---~ ;.-- ------~ rWI(nT) .1000.0 P(m) 2000 ""'" _. +1 ~ +2 ·""1 I .... 1-1------....:....--'------1 ~ .. (Ded·G 2{"") I!U';;"--)nBFI .... '-~_- _ .... - '" 1.~'31s:z.3llIIii,5J54zi:iiiJ"=IC Figure 4.1.7. Example #3 ofalternative model geometries producing equivalent observed responses A Geophysical Investigation of the Derwent Estuary 55 Chapter 4 Magnetic Modeling D~ __-- Jjloh »00'1"-_ - \...;.JdPC" r # '000» i ~ ~~-=---=-=------/ "WI(nT) .1000.0 I'(~) :lOOO 4000 __+1 3 , ' I .,.~., .' =1 rn,11' 1,1\, i" \1 1 1i I 1U;" f17 Oh, :o:~) ·100 I,::: ;:, " ". 1 .,..... ~rt ~ l909.87 !F:'314Cl45 Xi 537718Jllrri-~TS:-l Figure 4.1.8. Example #4 ofalternative model geometries producing equivalent observed responses Note that no serious attempt has been made to fit the data over the 0-2000m interval some sources influencing this section of the data probably lie off section. All of these models are geologically possible, although some are perhaps more likely than others. Although some of the solutions are similar in general body form, they differ in terms of depths, thicknesses and detailed shapes. Finally, although I have been emphasising that a range of possible solutions exists for any particular potential field model, I still believe that forward modeling is a useful interpretive tool. Although the models shown above do differ in detail, the gross form of the modelled geometries are reasonably similar, particularly in lateral extent. Since this modeling is used to help with geological interpretation, the lateral extents of bodies is important since the geological map is presented in plan view. The models also enable us to make estimates regarding the range of potential depths-to-top, and are useful in terms of defming fault or contact positions and possibly fault dips etc. A Geophysical Investigation of the Derwent Estuary 56 Chapter 4 Magr r Modeling As I have been emphasising, the magnetic models are poorly constrained and should be taken as indicative only. I have used estimates from these models to suggest the depths of r. some interpreted magnetic bodies on the geological interpretation (see chapter 9), but these depth estimates are based on assumed body properties and therefore could vary significantly (if the actual body properties are different to the modelled properties). They do, however, provide relative depth estimates for different bodies because the bodies are r generally modelled with the same susceptibilities and basic forms. This enables some sense of relative block displacement (ie. fault movement) to be gained. (' Most of the models I have shown have a good fit between the calculated and observed data - this does not imply they are correct, even if they appear plausible geologically. The models that do not have a good fit are from areas where the assumptions used in the modeling break down, ie. from where the geology cannot be approximated by two or two and-a-half dimensional models. In these cases, sources that have margins just off section or are entirely off section influence the observed response. On selected sections I have modelled bodies that occur off section - these bodies are shown in solid colours. I did not perform any regional/residual extraction prior to modeling. I cannot determine the distribution of the non-magnetic lithologies from my models and in general I do not know this independently, so it would be misleading to include sediments (~ or sedimentary rocks on my modelled sections. For this reason, I have only shown the magnetic bodies on the model profiles. I have added some approximate fault positions and displacement sense indicators on the profiles. In terms of general presentation, dolerite is r shown in orange and basalt in grey. Apart from that, the sections really require no further explanation - they are all very simple. The inducing field for all model profiles has a intensity of 62000nT, inclination of r. _72° and declination ofol4. Note that traverses 13 and 7 are not shown, although they do appear on the location map. r- r- A Geophysical Investigation of the Derwent Estuary 57 'r, !:1~ Cl.Eils Qbserlllltlons Model CalcUatlon IDols ,~,.i8W I'!(jndow ._ ·r.n~1 ~~I~.I"""7TIJ":'I"?1rA' I~ c.....=-==-:':===~.."..=""""~====-======,,--;--;-----===..- n..1_~lr-l...... ll"""'" 0.0 ·80.0 TMI (nT) ·160.0 i , , , i , , , , - i P(m) 0 2000 4000 ~ __ .9_ ·200 ,. ] ~\ ,L ,2 use=. \ \ ? I ·600 I I' .' '.' '. '. .... _~ ~l ~._~~' nn1~ ~lIoe;:~= -f ..a--- ... \ Az= 7~. Z(m) \ --- .. '1'. ·800 Ul 00 For H~, press FI Traverse 1. Figure 4.2 -Pj"IlI"O' 80.0 0.0 TMI (nT) ·80.0 P(m) o 2000 4000 6000 01 ~ I -- ,~tuse = 0.01800 i i III ·200 I I I 3 Suse = 0.01800 ·400 I I '. + ·600 I I. I I I II ' . _, 1 \ A:z = Z(m) 76.'ldog VI ·800 I ' I I A lit I ~ \D For H~, press Ft I 'I , I " ., Traverse 2. Figure 4.3 09 !f I ... !!; ~ro ~ ,:., ~ 00 3 0 s- o 0 ~ 0 8 i!5 '" 0 2- 0 0 ~ p b 0 b b ~ :?l I I t l + o '!" 1 ---+- I ~ ~ ~ 0" ~ b 00 - -.=:::: 8 I· IV 8 -:!:r o "" r-:--- r- I I I 1 r. f' I I ~ . - J -' ! ~ 1 I- 'N ~ g: - ... i -4 F-''""" w :!' r, -w . b 0 w, b .~ e 0 I':' . i I'~ ~ ,~ . q I IJ " ~"- -, E;JEile 80.0 0.0 TMI (nT) ·80.0 P(m) o 2000 'lOO0 iI •. _.01 ) --+ Su.e = 0.01800 I • 11 ,200 I ",\ / +Su.e. 0.01800' _---;- . -_,~ '- ·400 1 \ t ' '.,...... •• I, -6001 '.\..Lt 1 Az= Z(m) 77.0dtg ·800 I I ! I I 0\ ]X:53.~449.4€!V: - For H~.LJ:!8S5 FI IPi 3406.CJ;UJF: 141.79 l:!m032';j Traverse 4. Figure 4.5 Ori ~ Ble t!.eIp ~ 320.0 ~ ...... -_..."'...... -• 0.0 - TMI (nT) -320.0 P(m) o 2000 4000 o I I 3 •+ Susc = 0.06000 -2, 1 __ • __ 1Susc· 0.01800 :-,-;- ~ ~ -400 I \ S' ,~ . . . I -600 I I f';za 77.6deg Z(m) -800 J I • • , 0 • I. 0\ ,--,~, ~ N For Help, press Ft \/ ,I, ~ ,\" I. .•.. Traverse 5. Figure 4.6 .... '\ ~w tletJ 200.0 0.0 ~flll;"iI.'ItI~1 ~ ~ TMI (nT) -200.0 2000 4000 o Ol~ =~~~"sc~_oo~18DOc=~=~==~~~~~~~===~~L=_I -flrul. -. • ~~"'""r-r""'j 1 ~~ -200 I \ Suso E 0.02500 I' • ,+ I I .'+ .~ -'lOO I ''''.Lt . I I -eooI I. .- J I Z(m) Az =77.0deg r ·800 I : !.. -, .. 0\ For Help, press Fl "" ,j ....., --. v.l Traverse 6. Figure 4.7 800.0 0.0 Tlull (nT) t ·800.0 P(m) 0 2000 4000 5000 0 ~~--...~u.c = 0.07500 ...... , ·200 .. '~ ...... :.::...... __._.~ -_.. --.: -400 ·500 Z(m) Az= 77.2d.g ·800 0'1 ..j::.. For H~, press Fl J(: 533128.0LI'1 5241s65}. Traverse 8. Figure 4.8 ''0 ~ew !::!eIp I5l Eile Qbseryatlons Model Calc{jatlon Tools '_~\'!indow . __._ .., ---e.:: 1-_1 j~...... -r __ I __ J i I_-It 320.0 \ 0.0 ~l~ TMI (nT) -320.0 P(m) o 2000 4000 6000 o ...... ...... -. ~ ~ 1: +- .. - . , \ + Suse = 0.03000 -r ...-",, -200 I .. 1 .I ~+ --.----~.- , . ,- .. --+ -400 I / ...... -, , i . i " --~'--- : '/~-' r I .. :' -600 ! i :/ i· . '/ !':z = 76.~d.g Z(m) [ f I ,.r I -800 -. .. I lJI0\ ForH~,__ pr - (P: 5480.00 IJF:5.54.4200j(X:533318.';JV: 524368l.tf Traverse 9. Figure 4.9 ) f t 160.0 0.0 TMI (nT) ·160.0 P(m) a 2000 4000 a I 1 11 t S use = 0.03000 ~ ·t ~use 0.01S00 Rom H• 0.00 .Rem H ~ 0.00 ~- 0 ~ -~ ·200 I . . \'...." ~~ ,- •.- . ",3Suse = 0.06000 to I Rom H = 0.00 .400 I 1, \ • Iv'l ~Zci'03000Rom H• 3.00 ~I~'~"'''''''''d i:."'.' ~ "I ':' vv>,\ •I I ~aal " AI. = 76.Qdeg Z(m) 0\ ·SOO I I 0\ For HOP,press FI I, - -- I TTTTJ 'I Traverse 10. Figure 4.10 "'I ...... -= 15lEjle Qbserva_oonsModel CakLJaoon Tools :!Jew~ ..__.. ~~ I~ ~ ~r=---.,_ 1 j i l_ -- • 800.0 0.0 TMI (nT) _ ,rlf,lr'ill"L!lIIIItIf.'~H''''_llffill\.~I!+tJ,r~ ~llilIUlIiIIJlll"'i ·800.0 P(m) o 2000 'lO00 d 0, 1 ~~~~~.g7gg0 rVHVV~V;";";~~11 ...... >?·.~"V~·,_ ,,_.,,=~;Y~=:;;t~~~~llllllljll!W, ~I "rn'-'" I v~.: ~ ~_ ~ ~:_~ • Rt!m H· 0.00 . \'f ~.~..~. ';' .\ \. Y '/ ., y '/ \' ~v'.'- ,'-", Zl:m) jtig::::l~1':,\ ...... '--....__ Of \' of '; " t '... " ',' r 'vi y"~' ',. 'y' '1 \ .800 I I I \,"'''''V'IY 'i 'r' °r'• If 'J t I ! ! Ij CJ\ 52~2591.f For H~,pre5S 1:1 - -. - .----- -.... ----- ../P:362o,531jF,lI61z.93:f&]:lIB65.llE'!I': = -....l . . Traverse 11. Figure 4.11 hr·... Cl Ble Qbservations Model ~alciJatlon IDols yrew ..:c=:r;::;..:~--T~- -'::--i~:"-t i,. i..--... 80.0 0.0 TMI (nT) ·80.0 P(m) o 2000 4000 dE· 0, 2 ~3 .200 I lie" .. "" '"'''' _ ..",-:.-:-:5:-.;_JSuso= 0.05000 ~",::'7"S.~1.'-- 1 - +Suso= 0.01800 • I _~." ·400 I ~ I . __ I' ~ool I I Pa= 72.7dog Z(m) 0\ -800' I , I • I' ~-~~" 00 ForH~,pressFl I.. " ' Traverse 12. Figure 4.12 '.. _..'Ix Qbservaliol'6 Model£aktJation 10015 yjew :t£indowl:i~ _,"I X 320.0 0.0 TMI CnT) -320.0 P(m) o 2000 "1000 °r------~:=:::-:-J:::::-----:~~;:------.:::=------.:~~------; i: ~USc - 0.03000 ...... I .' 2 + Susc = 0.03000 " I I r""""""'-+SUS.C=0.03000 / I I . ·200 I • •.,l', I r I' 3. , :'.' .---, .....--..;.- ... ---" .. ·400 I • I ·600 I I Z(m) Az.·125.5dtg 0\ -800 I I " '! \0 tCJrH~J P':8SS Fl ,! i__ ._. -.f:' •• __ • .... --_. l .. -_.- - .! j Traverse 14. Figure 4.13 Chapler4 Magnetic Modeling Modelled Magnetic Traverse Locations 5250000 5245000 5240000 5235000 5230000" I K t ( t t ~5230000 525000 530000 535000 540000 2500 o 2500 Me\3rs Slmpldi od Geology • Cl OJal:errury Cl Terbary N Cl Triassic On-shore Geology Polygons Cl Panni Figure 4.14. Location map for modelled magnetic traverses. Although the locations of traverses 13 and 7 are shown here, the modelled sections have been omitted. A Geophysical Investigation of the Derwent Estuary 70 Chapter 5 Seismic Survey c~ CHAPTERS SEISMIC SURVEY r The seismic source used for the Derwent Estuary work was an EG&G Geometrics Uniboom "boomer" operating at 200J and 4Hz. As with all types of seismic reflection surveying, source selection is a trade off between resolution and depth of investigation. Depth of investigation is dependent on both the properties of the materials under (' investigation and the frequency content and amplitude of the source pulse. Higher amplitudes and lower frequencies provide greater depth of investigation, but at the cost of resolution. High frequencies provide better resolution, but with reduced depth of (-, investigation. Chirper systems are designed to get around this limitation by using a linear frequency modulation sweep (eg. from 2 kHz to 16 kHz), but these systems are less readily available at low cost. The survey specifications (and budget) are typically the controlling factors in the choice of source. ('. The boomer plate was mounted in a catamaran and powered by a 6.9kVA petrol generator. The boomer source consists of a coil of wire mounted in a rigid, non-conducting medium r (eg. epoxy resin) that sits flush with an aluminium plate. A high voltage discharge from the power source energises the coil, and the magnetic field associated with the electrical flow in the coil induces eddy currents in the aluminium plate. The eddy currents induce magnetic fields which oppose those produced in the coil, and the aluminium plate sharply ". pushes away from the coil - producing a short, broad spectrum pressure pulse. The seismic equipment was hired from the University of Sydney, and operated by Mr. David Mitchell (the seismic technician from USyd). " r. c' A Geophysical Investigation of the Derwent Estuary 71 Chapter 5 Seismic Survey Figure 5.1. The boomer catamaran (the plate is beneath the wooden cover). The catamaran goes up the other way in the water. The receiver· array consisted of a hydrophone streamer (or "eel") containing ten hydrophones wired in series, inputting to a single channel analogue recorder with output onto electrostatic paper. The paper record has time (ie. along line time) on the horizontal axis, and two-way time (a proxy for depth) on the vertical axis. The source-receiver offset was around 4 metres. Time varying gain and uniform 'trace' amplification with clipping were applied prior to output. A bandpass filter with a low cut frequency of 700Hz or 1.5kHz and a high cut frequency of 5kHz was also applied, with the low frequency cut off being changed as required. Navigation was via Ashtech and Garmin 48 GPS systems, without differential corrections. Seismic Terms At this point it is worth introducing some terms for those not used to seismic data. In terms of seismic profiling, a reflection is a return of transmitted energy from a physical boundary within the earth. A reflection can only occur at a boundary between two substances with differing acoustic impedances. The acoustic impedance (AI) of a material is the product of the sonic velocity (the speed at which the sonic pulse travels through the material) and the density of the material. The proportion of energy transmitted and reflected at any particular boundary depends on the AI contrast across the boundary. More energy will be reflected if there is a large AI contrast than if there is a small AI contrast. We can quantify this using A Geophysical Investigation of the Derwent Estuary 72 Chapter 5 Seismic Survey the reflection coefficient and the transmission coefficient. These coefficients give the proportions of energy reflected and transmitted (respectively) at a particular boundary. Figure 5.2. The hydrophone streamer (eel) trailing behind the boat. A bamboo outrigger was used to move the streamer out past the boat wash. The paper record was regularly time stamped, and positions were referenced from the GPS logs (see appendix). With the system operating at 4Hz, the maximum recorded two way time was 0.25 seconds, but typically the record is interpretable for a maximum of around 0.125 seconds. A Geophysical Investigation of the Derwent Estuary 73 Chapter 5 Seismic Survey /._ r,~- Figure 5.3. The electrostatic printer. The seismic survey lines were not pre-planned in detail, and are not on a grid as such. See the trackpaths map for traverse locations. The vessel used for the main seismic survey in 2001 was the FMV Mallanna, an aluminium sharkcat approximately 16 feet long. A preliminary seismic survey using the same equipment (but a slightly smaller boat) was carried out over 3 days in August 2000. The trackpaths map shows the locations of the traverses for this trial survey. A small two-boat survey was also carried out at the Bridgewater mudflats prior to the commencement of the 2001 survey - although I have not interpreted any of the data collected near Bridgewater, I have shown the line locations on the trackpaths map. This survey was conducted as part of the proposed gas pipeline site investigation. Two boats were required due to the very shallow water at this locality - each boat carried half of the equipment. A Geophysical Investigation of the Derwent Estuary 74 Chapter 5 Seismic Survey ,-.. Figure 5.4. The FMV Mallanna, with Mr. David Mitchell on board. Figure 5.5. The two boats usedfor the Bridgewater survey. A Geophysical Investigation of the Derwent Estuary 75 Chapter 5 Seismic Survey Seismic Survey Trackpaths 510000 515000 525000 530000 545000 5270000-1 + + + + 1-5270000 + + 5265000 5260000 5255000 + 1-5250000 524500 5245000 5240000 + 1-5235000 5230000-1 + + + + 1-5230000 5225000-1 + + + + + 1-5225000 510000 515000 520000 525000 530000 535000 540000 545000 2500 0 25005000 Meters Major Roads N N ~ N Seismic Paths 2001 Coast from 1:250,000 digital coverage Major Roads from 1:500,000 digital coverage N Seismic Paths 2000 A Grid: Australia Map Grid Zone 55 Datum: Australian Geodetic Datum 1966 Figure 5.6. Seismic trackpaths map showing locations of all seismic reflection profiles (2001 and 2000 surveys). A Geophysical Investigation of the Derwent Estuary 76 Chapter 5 Seismic Survey (~ Acoustic Turbidity , It quickly became apparent during data acquisition that a large portion of the Derwent Estuary was characterised by a shallow subsurface with acoustic properties that we had not expected. Rather than seeing clear, coherent reflectors at most locations, we observed expansive zones of chaotic reflections (noise-like) that appear as dark smears on the r: seismic record. These zones tend to have sharp lateral.boundaries, so the transition from coherent reflections at large two-way times to loss of detail is essentially instantaneous. These zones typically appear to extend to the seafloor, but sometimes have a thin veneer of sediment overlying them. Rarely, they terminate vertically at a horizontal reflector. The n vertical extent of these zones is difficult to judge in almost all cases - it is exceptionally rare to find a coherent reflector beneath them. This could mean one of two things - the zones could be quite thick, or they could be relatively thin but their acoustic properties \' prevent the imaging ofthe underlying reflectors. I consider the latter scenario more likely, for reasons that will become apparent. The Basalt Theory n As mentioned in chapter 2, there are a number of basaltic volcanic centres known around Hobart including some bordering the Derwent Estuary (eg. at Blinking Billy Point in Sandy Bay). It was initially suggested that the zones of chaotic seismic response could be ('> due to the presence of basaltic lavas and associated rocks within the stratigraphy. A large acoustic impedance contrast would exist between hard basalts and soft sediments, and it was thought that this could be sufficient to prevent the propagation of the seismic signal -I) through the basalt (ie. it would all be reflected off the upper surface of a flow). This suggestion has a reasonable theoretical base, and would explain the inability of our seismic system to image reflectors beneath the upper surface of the zones of chaotic response. It was also favoured because a general co-incidence between the areas of unusual seismic (') response and regions of anomalous magnetic intensity had been noted. This correlation was not perfect, but several factors combine to suggest that it need not be. Firstly, some of the Tertiary basalts are effectively non-magnetic. Secondly, a thin sheet of magnetically \) susceptible basalt might be enough to degrade the seismic signal but not be detectable by the magnetometer with sensor heights typically greater than I5m above the seafloor. Forward modeling confirms that with a susceptibility of 0.02 SI units, a 5m thick sheet of " A Geophysicallnve:stigation ofthe Derwent Estuary 77 Chap[er 5 Seismi c Survey basalt at ISm depth (with an inducing field of 62000nT, declination 14° and inclination 72°) produces negligible magnetic response, assuming thinning margins. Finally, where the magnetic field is anomalous but the seismic response is not, we could argue that the magnetic sources are buried below the depth of investigation of the seismic signal. :r;l""j;.~~ ~ ~ ~ ~ ~ ..o1if~ fDI~lnlii[j ,alifIr-=FI .... I~Rf~;I~I.~lI . /' 400.0 200.0 0.0 - = -200.0 ThlI(nT) -400.0 P(m) 400 800 1200 o,I-----,----~~~~~~~-~~~~~--~-~-----~------.., ~lf------______1 -101f------~ -151 ,1 seu_.'c~,~:~2l1~O,_ .. '~"" -1201 :~:: ..._:: :::: ::;:: :: ::,".: ":.~,-, ~1r------______1 ~ =O.Od81i1 Z(m) ~,":;l.L-~-~-~--'------~-_~_~--.-=_~-._------'---~-~-~-----";;'="~.~-;:-;~~..~7.,."ijjll'i~~~•. ii~.::;'l~)I,~.. These arguments do not explain some features associated with the anomalous seIsmIC response. Firstly, we observed that it typically occurs in the deepest portions of the river. Secondly, the upper surface of the anomalous zone is always restricted to the shallow subsurface - it does not occur deep within the stratigraphy. Thirdly, the lateral margins typically have coherent reflectors dipping into them - suggesting that whatever causes the anomalous seismic response was emplaced into low relief areas of the palaeotopography. These factors can be explained in terms of the basalt theory using the following arguments. High temperature, fluidal basalt lavas typically flow freely under the influence of gravity, and are often emplaced along topographic lows. This explains why we see the 'basalt' in channels, and may explain why it occurs in the deepest portion of the river - ie. perhaps the modern channel profile records TertiarylPleistocene relief. Alternatively, erosion could A Geophysical Investigation or the Derwent Estuary 78 Chapter 5 Seismic Survey have removed poorly consolidated post-basaltic cover to expose the more resistant lavas this could also explain the first observation. ". Disproving the Basalt Theory These arguments are all reasonable and plausible geologically, but are not necessarily correct. To answer the question regarding basalt, it was necessary to consult the core logs. " As mentioned, no deep coring has been performed in the Derwent Estuary except as part of the studies regarding the two bridges - the Bowen Bridge and the Tasman Bridge. Although these were outside the principal study area, seismic profiles were recorded , alongside both of these bridges. The profiles from both bridge sites contain zones of the anomalous seismic signal. At the Bowen Bridge site, basalt occurs adjacent to both abutments. However, no basalt was intersected in any ofthe bore holes. At the Tasman Bridge site, basalt was intersected in over half of the bore holes, and occurs beneath the zone of anomalous seismic response. However, the basalt at the Tasman Bridge occurs deeper in the stratigraphy - well beneath the upper surface ofthe acoustically anomalous zone. In both cases, the anomalous seismic response begins at a two wiry time that places it within the youngest, fine grained sediments of the section. Clearly, the anomalous seismic response that had been observed was not due to basalt within the stratigraphy. Furthermore, although the youngest muds extend right across both sections (with laterally varying thicknesses), the anomalous acoustic response does not. Yet no change. in sediment characteristics is noted on either section at or near the boundaries of the anomalous seismic response. This suggests that the property of the sediments causing the seismic signal degradation was not apparent to the core loggers. ,". The Evidence for Shallow Gas Obviously, this raised the question: what property of the fine grained sediments might be causing the seismic response that we had seen? One clue came from Colhoun and Moon " (1984), who noted that at the Bowen Bridge site the uppermost formation of massive estuarine clays is "rich in colloidal organic matter (10-22.5%)" (page 224), suggesting an organic process could be responsible. After studying the literature and the data, I have r formed the opinion that the observed seismic response of the acoustically chaotic portions of the Derwent Estuary can be explained by the presence of disseminated gas bubbles ,. A Geophysicallnvestigation of the Derwent Estuary 79 Chapter 5 Seismic Survey within the sediments. The seismic evidence for the presence of gas is compelling, and is discussed below. .~ I cl r f " Fh Line J I' : :'~.:' Figure 5.8. Example ofa gas blanketIram the literature (Taylor, 1992). We can obtain a significant body of evidence for the presence of gas in shallow sediments from high resolution seismic data, and we can break this into two kinds of evidence "direct" and "indirect" (Judd and Hovland, 1992). Direct evidence is the acoustic expression of physical phenomena associated with gas, and indirect evidence is the morphological and/or structural expression of the presence (or former presence) of gas. Firstly, I will consider the direct evidence. A Geophysical Investigation of the Derwent Estuary 80 Chapter 5 Seismic Survey Acoustic Turbidity Distribution 530000 526000 + + + 5260000 525500 + + + 1-5255000 525000 + + 5250000 524500 + + + ~5245000 524000 + + i ("-.J +/ J I + r + 1-5240000 523500 + + 1-5235000 \\ 520000 525000 530000 535000 540000 2500 o 2500 5000 Meters N I o Acoustic TUrbidity Coast from 1:250,000 digital coverage A Grid: Australia Map Grid Zone 55 Datum: Australian Geodetic Datum 1966 Figure 5.9. Map of acoustic turbidity distribution as defined from the high resolution seismic profiles (for 2000 and 2001). A Geophysical Investigation of the Derwent Estuary 81 Chapter 5 Seismie Survey r' Acoustic Turbidity Gas in shallow marine sediments can be expressed in a number of ways. One of the most common is as acoustic turbidity on high resolution seismic profiles. Acoustic turbidity is r defined by Judd and Hovland (1992) as the scattering ofacoustic energy by disseminated gas in fine grained sediments, recorded as dark smears on the seismic profile. Taylor (1992) and Yuan et al. (1992) define acoustic turbidity in a similar way, but extend the (. definition to include absorption effects as well as scattering. Bouma et al. (1987) do not use the term acoustic turbidity but refer to acoustical voids or wipe-out zones that they attribute to bubble phase gas occurrences in the shallow sub-bottom. ~ Acoustically turbid zones have been classified according to their morphology and other characteristics into three types (blankets, curtains and plumes) by Taylor (1992). The descriptions and examples of acoustic turbidity from the literature are strikingly similar to .-, the acoustic response observed in the Derwent Estuary. The nature of the response (noise like, but wit!) higher amplitudes) is more satisfactorily explained by scattering and attenuation by gas than by a strong reflection from the surface of a lava flow. Both "curtains" and "blankets" occur within my data, although blankets are more common. The distribution of this acoustic turbidity is shown on figure 5.9. Enhanced Reflections (\ Gas can also be manifested as enhanced reflections. These are the equivalent of bright spots in petroleum exploration seismic data, and result from the accumulation of gas in porous horizons ego in a sand layer beneath an impermeable clay. The accumulation of gas has the effect of lowering the acoustic impendence of any given sediment type compared (' to water saturated sediment, enhancing the reflection from the boundaries of the gas charged horizon. It is common to find enhanced reflectors extending from the margins of acoustically turbid zones (Judd and Hovland, 1992). ,.... u There are enhanced reflections present within my data, particularly extending into the edges of acoustically turbid zones. These reflections occur at the upper surface of the gas charged zones and I interpret them to indicate an increase in reflection coefficient due to a o decrease in the acoustic impedance of silts caused by gas build up. n A Geophysical Investigation of the Derwent Estuary 82 Chapter 5 Seismic Survey r Acoustic Blanking Acoustic blanking is a phenomenon characterised by faint reflections, or an absence of " reflections on a record where gas is inferred to be present (Judd and Hovland, 1992). Blanking by itself is not sufficient to demonstrate the presence of gas, but in combination with other evidence such as acoustic turbidity, it can be used to infer the presence of gas. It c', can result from the mixing (homogenisation) of sediments by migration of pore fluid (ie. gas) or by absorption without scattering in overlying gas rich sediments (Judd and Hovland, 1992). Acoustic blanking does occur in my data, but as I have mentioned this cannot be unequivocally attributed to the presence of gas. " Phase Reversals If gas is present within a package of sediments, a phase reversal may be evident in the seismic data. This occurs when the energy pulse is reflected from an interface where the upper layer hl\s a higher acoustic impedance than the lower layer (note that in the general case, acoustic impedance will increase with depth). Of course, gas charged sediment will tend to have a very low acoustic impedance (since gas will tend to lower both the bulk velocity and density of any given sediment type), so we could reasonably expect to see seismic phase reversals associated with gas. Phase reversals do appear to be present on some of my seismic profiles, but it is difficult to be certain in the interpretation of these (' features. Taylor (1992) states (in regard to 'blankets' of acoustic turbidity) that there is "...rarely any evidence of phase reversal at the upper surface of the feature." (page 1140). Furthermore, attempting to pick a phase reversal from an analogue record can be difficult (James Tayton, pers. comm.). Nevertheless, where the acoustic turbidity has clearly (; defined margins and does not extend quite to the riverbed, I believe phase reversals can sometimes be detected. () ("\ n A Geophysical Investigation ofthe Derwent Estuary 83 Chapter 5 Seismic Survey Figure 5.10. Phase reversal example 1. See notes below. Possible Phase Reversals Figure 5.11. These two small sections show possible phase reversals associated with acoustic turbidity. The white/black reflector change at the indicated positions is suggestive ofa phase reversal (the analogue record represents different signal polarities with different shades). Note the 'trapping' of the acoustic turbidity beneath faint horizontal reflectors in these sections. In detail, the acoustic turbidity is marked on the upper surface by a strong reflection (black). This strong black line passes laterally into a white line approximately in the positions of the arrows. This signifies a lateral change in signal polarity, which in these cases is clearly associated with the edge ofthe acoustic turbidity. This may represent a phase reversal. A Geophysical Investigation of the DeIWent Estuary 84 Chapter 5 Scismie Survey r Gas Seeps Seafloor gas seeps can be detected acoustically. These occur as dark smears witbinthe water column on a high resolution seismic record (shoals of fish have a similar acoustic I' expression). Field and Jennings (1987) studied sediment destabilisation associated with earthquakes and provide example of seepages visible on high resolution seismic reflection profiles. Bouma et al. (1987) state that "Gas escaping through the water column can be clearly seen as vertical "V's" on high frequency records such as 12kHz." (page 74). (\ Numerous water column disturbances occur on the Derwent Estuary profiles that are not associated with speed, direction or gain changes. These water column disturbances show a strong spatial relationship with the zones of acoustic turbidity (see figure 5.15). At least I' some of these could be gas seeps. Velocity Pulldown ,. A final physical phenomena associated with the presence of gas is velocity pulldown. Dobrin (1976) explains it in the following way: "The low velocity of gas saturated sand...gives rise to lateral anomalies...in the time required for seismic waves to pass through the sand." (page 348). It is recorded as an increase in two way time to all reflectors beneath a layer with elevated gas concentrations. Since variation in gas concentration affects bulk sediment sonic velocities, a lateral increase in gas concentration within a given layer will result in an increase in two way time for reflections from below the upper surface of the gas rich layer. There are excellent examples of pulldown on the Derw.ent r Estuary profiles. Indirect Evidence (' Indirect evidence of the presence or former presence of gas comes from sedimentological and/or structural features that result from the migration or accumulation of gas. If observed on seismic profiles, the reflectors typically do not have unusual acoustic characteristics and n could easily be interpreted as being formed by other processes if the interpreter did not realise gas might be present. Indeed, the interpretation of these structures is not always unambiguous even if the presence of gas is suspected or known. n I' A Geophysical Investigation ofthe DeIWent Estuary 85 Chapter 5 Seismic Survey Figure 5.12. This is an example ofvelocity pulldown and seepage. The dark targets in the water column are probably gas seeps (see text). The signal associated with the seepage is weakest at early two way times - this is due to the gas dispersing as it rises from the seafloor. The position of a curtain of acoustic turbidity is indicated by A. Reflectors Band C appear at earlier two way times away from the curtain, and appear at greater two way times directly beneath the curtain. This does not represent a change in the physical depth of the interfaces imaged as reflectors Band C, rather it is a result of a lateral changes in gas concentration within the sediments. The gas concentration is greatest in the position of the curtain, so the bulk sediment sonic velocity is lowest in this position. This causes reflectors to be imaged at greater two way time beneath the acoustic turbidity curtain. Also see text for discussion ofpulldown. A Geophysical Investigation of the Derwent Estuary 86 Chap'er 5 Seismic Survey I' Pockmarks are depressions on the seafloor caused by the escape of a pore fluid - in most cases this pore fluid is gas. The pockmark is a result of the escaping gas removing the soft sediments (Hovland, 1982). The formation of pockmarks has been artificially induced by r blow outs during drilling (Judd and Hovland, 1992). Pockmarks may be associated with other features, such as columnar disturbances. Although not abundant in my data, there are a few features that appear to be pockmarks, based on the examples reported in the (' literature. Seabed domes may predate the formation of pockmarks - they are broad, low relief features possibly caused by the displacement ofwater by gas in pore spaces, resulting in an r~· inflation effect. They have been associated with gas in several studies (Judd and Hovland, 1992). Columnar disturbances are columns of breached or disturbed reflectors, where the disturbances are attributed to the vertical migration of gas - these may occur directly r beneath pockmarks (Hovland, 1982). Mud diapirs are bodies of plastically deforming, gas rich clay or II)ud that rise buoyantly and may penetrate the seafloor (Judd and Hovland, 1992). Both of these features are caused by the buoyancy of gas and assisted by low confining pressures of the shallow marine environment, and can be recognised on seismic profiles. There is no evidence for seabed doming, but there are disrupted reflectors where the cause ofthe disruption is not clear on the Derwent Estuary profiles. ( r " n r A Geophysicallnvestigation of the Derwent Estuary 87 Chapter 5 Seismic Survey Figure 5.13. This is an example ofa pock mark (see text) from the Derwent Estuary. ;: I -I 290 01 -IQ' r' Figure 5.14. This is an example ofa pock mark and gas curtain from the literature (Taylor, 1992). A Geophysical Investigation of the DeIWent EstualY 88 Chapter 5 Seismic Survey (' Clearly, the evidence for gas is strong based on the seismic record. Several features that are strongly suggestive of the presence of disseminated gas are definitely present within the data, and others are arguably present. But why does this gas affect the seismic signal? r The attenuation and scattering effects ofacoustic signals by bubbles are complex, and only partially understood. Some of the significant factors include: occurrence (bubbles can " occur wholly within the pore water between grains, they can displace grains, or they can enclose grains of sediment); bulk sediment shear modulus; bubble size distribution; bubble resonant frequency and acoustic frequency (Wilkens and Richardson, 1998). Suffice to say that the presence of gas bubbles can result in very rapid signal attenuation, with (frequency (' dependant) acoustic attenuation of hundreds to thousands ofdB per metre possible for gas charged sedirnents. (~, These recent observations can also be used to help explain and reinforce the long-held view of Leam.an (pers. comrn., 2001) that the uppermost layer of silts within the Derwent Estuary have low acoustic velocities. The first suggestions that this was the case came during refraction seismic surveying as part of the site investigation for the second Derwent crossing study at Dowsings Point. During this survey, Leaman (1975b) deduced that the silt velocity was approximately 1050m1s. Note that the velocity of seawater is around 1500 m1s, and water saturated sediments would be expected to have velocities equal to or greater than that of seawater. The presence of free gas as bubbles within the sediment has the effect of lowering the bulk velocity significantly - Taylor (1992) states that as little as 0.1% gas (by volume) can lower bulk velocity by up to one third when compared with bubble free, water saturated sediment. r· So then, what sort of gas are we are talking about, and where is it coming from? Several lines of circumstantial evidence suggest that if gas is present, it is likely to be biogenic in (' origin and being produced in situ by bacteria. r r A Geophysical Investigation ofthe Derwent Estuary 89 Chapter 5 Seismic Survey Water Column Disturbances 530000 526000 + + + 5260000 525500 + + + 5255000 525000 + + 5250000 524500 + + + 5245000 524000 + + + 5240000 523500 + 5235000 o \) 520000 525000 530000 535000 540000 2500 o 2500 5000 Meters N I - ! • Water Column Disturbances o Acoustic TUrbidity Coast from 1:250,000 digital coverage A Grid: Australia Map Grid Zone 55 Datum: Australian Geodetic Datum 1966 Figure 5.15. Map of water colum disturbances as defined from seismic profiles. Note correlation between water column disturbance distribution and acoustic turbidity distribution (see text for discussion). Disturbances known to be side echoes etc are not shown. A Geophysical Investigation of the Derwent Estuary 90 Chapter 5 Seismic Survey r As mentioned in the introduction, the modern Derwent River has deposited fluvio-deltaic mudflats near Bridgewater. At this location, the suspended fine grained sediment load of n the river is flocculated by the interaction with salt water, and the resulting muds are rich in organic material. Since sea level has fluctuated significantly throughout the Pleistocene (down to around -125m below present twice in the past 250ky), deposition of similar material is likely to have occurred along most of the length of the (modern) Estuary at " some point in the last 250,000 years (see appendix I), providing a source of organic matter for biogenic gas production. " Secondly, there is a remarkable correlation between the bathymetry of the Estuary and the occurrence of the acoustic turbidity - the acoustic turbidity is confined to the deeper portions of the Estuary, particularly south of the Tasman Bridge. While this could be due to confining pressure associated with the deeper water retaining more gas within the " sediments, it is more likely that it indicates the area of maximum gas production. This in turn is linked to the first point, regarding the accumulated mass of organic rich sediments (ie. the gas is being produced from channel fill sediments, deposited during periods of 0. lower sea level - these occupy the deepest parts of the bathymetry). Okyar and Ediger (1999) noted a similar correlation between water depth and acoustic turbidity distribution and concluded that this was a reflection ofsedimentation patterns. ('. Thirdly, we can compare the Derwent results to other shallow marine/estuarine sediments that display acoustic turbidity. Park et al. (1999) found that acoustic turbidity in the Korean Sea was likely to be due to the in situ production of gas in recent muds, Okyar and n Ediger (1999) came to a similar conclusion about acoustic turbidity in Pleistocene sediments of the Black Sea. Mullins et al. (1991) suggested the same cause for acoustic turbidity in the sediments of Lake McDonald in North America. Taylor (1992) states "... .it n would appear that there is a wealth of evidence for the presence of shallow gas in nearshore marine sediments around the U.K..", but he goes on to state that acoustic turbidity in the estuaries of the Thames, Mersey and Forth is generally not associated with local structures (implying that it is more likely to be biogenic than thermogenic). And as a " final point from the literature: "Most gas in surficial, shallow-water sediments is biogenic and originates from the generation of methane as a by-product of metabolism by methanogenic bacteria" (Richardson and Davis, 1998). \: A Geophysicallnvestigation of the Derwent Estuary 91 Chapter 5 Seismie Survey ,". Finally, there does not seem to be a close spatial association between the acoustic turbidity and known structures - or structures inferred from the magnetic survey data - in the Derwent Estuary. This suggests that the gas causing the acoustic turbidity has not migrated (' up from deep within the basin, and is unlikely to be thermogenic. Equally, the gassy sediment is found in the Derwent Estuary - in a graben. Certainly there are large, deep rooted structures near the acoustically turbid zones, so it is perhaps misleading to suggest r there is no spatial association with large structures. In sununation, there is no firm evidence regarding the origin of the gas in the shallow sediments of the Derwent Estuary. It is my personal opinion that it is probably biogenic, ,~ but I can not demonstrate this conclusively. Biogenic Methane Production r The term 'biogenic gas' is used to refer to methane produced by bacteria. This is quite different to thermogenic gas, which is gas produced by the thermally driven decomposition of organic material. The bacteria that can produce methane all belong to the Archrebacteria r group, and are unusual in that they are strictly anaerobic, albeit with varying toxic responses to oxygen (Floodgate and Judd, 1992). Although there are only a few species, the methanogens are relatively common in muds, soils, and the intestinal tracts of many r animals (especially ruminants). Unlike thermogenic gas, biogenic gas production does not require the deep burial of organic material and usually occurs in the shallow subsurface. In relatively impermeable muds, available oxygen is quickly depleted by other bacteria and anoxic conditions may be achievable within millimetres of the surface (Floodgate and (' Judd, 1992). The methanogens can only produce methane from a limited number of substrates, including acetate, methanol and hydrogen+carbon dioxide, so other bacteria are required to break down polymeric organic substances to these simpler compounds. The ,(' terminal electron acceptor from the chemical reactions varies depending on the environment (which influences the available ions), but in brackish or salty water sulphate can be very important - leading to the reduction of sulphate to sulfide, and the production of hydrogen sulfide. Since the interplay between species is complex, and environmental (' conditions vary greatly it is very difficult to predict exactly which gases will be present in r- AGeophysical Investigation of the Derwent Estuary 92 Chapter 5 Seismie Survey c' the muds of the Derwent Estuary, except to say that methane is likely to be most abundant, and some carbon dioxide, hydrogen and hydrogen sulfide may be present. (") Dissolved Water-Sediment Metabolic Processes Species (Biogeochemical zones) Air r> Photo synthesis O2 (photic zone) Aerobic Water Respiration (") so/ HS (i HC03- Sediment Anaerobic Respiration CH4 H2 (") Figure 5.16. Modified from Okyar and Ediger (1999) after Rice and Claypool (1981). Okyar and Ediger (1999) suggest that in cases where the upper surface ofan acoustical/y turbid zone paral/els the seaf/oor, the upper suiface may represent the boundary between the sulfate and carbonate reducing zones. If this is correct, then methane and hydrogen will only occur as dissolved species below the upper boundary of the acoustically turbid zone. (") The distribution of acoustic turbidity has been found to vary with oxygen availability on a seasonal basis in some locations. For example, Muller (1976) reported that acoustic turbidity (detected with a 30kHz echosounder, capable of perhaps Im sub-bottom (") penetration) became more widespread when the deep water became oxygen depleted. This is almost certainly linked to the anaerobic requirements of the methanogens. (j Biogenic methane production and the processes associated with it can lead to carbonate precipitation. Measurements of carbonate concentration in shallow cores taken from the Derwent Estuary indicate generally higher values towards Storm Bay, and a zone of elevated carbonate concentration extending up the centre of the Estuary. This zone of (') elevated carbonate concentration is broadly coincident with the mapped gas occurrences, ('") A Geophysical Investigation of the Derwent Estuary 93 Chapter 5 Seismic Survey ,~' and may in part reflect carbonate precipitation via processes associated with biogenic methane production, r The carbonate concentration map was kindly provided by Dr. Peter Harris from the Antarctic CRC at the University of Tasmania and is reproduced here with his permission. The suggested interpretation I have provided is my own, and in no way linked to Dr. " Harris. ('-, ~ (", r r " A Geophysical Investigation ofthe Derwent Estuary 94 Chapter 5 Seismic Survey 1~1~' "7,,l' ..n.v ~'II! W ...-- E ., ~ ~ -IJ' l~lA· Figure 5.17. Map ofcarbonate distribution in the Derwent Estuary, provided by Dr Peter Harris. There is a general correspondence between the zones of higher carbonate values (20-30%) and the acoustic turbidity distribution. See text for discussion. A Geophysical Investigation of the Derwent Estuary 95 Chapter 5 Seismic Survey ,-- Simple chemical arguments dictate that if a gas is present as bubbles within wet sediment, then the aqueous interstitial fluid must be saturated with that gas. This deduction led to the development of the headspace analysis technique of Kvenvolden and McDonald (1986). '---, The technique involves sealing a wet sediment sample in a gas tight vessel, heating to liberate dissolved gases, and analysing the "headspace" gas (i.e. the gas in the container a mixture of air, water and any gases released from the sediment). A slightly modified !" version of this technique was applied to the sediment samples taken from the Derwent Estuary in an attempt to determine whether or not gas is present in the shallow sediments (as the seismic data suggests). ,r. Sediment Sampling Samples were collected along a single transect, from Cartwright Point on the western shore ,--- to just south ofTranmere Point on the eastern shore. This transect was chosen based on the seismic data, since the acoustic turbidity on the seismic profiles appeared to extend to the riverbed (to the resolution of the seismic system) in this area. The wet sediment samples were collected using the Craic corer belonging to TAFI. This corer has a rope that is used both for deployment/retrieval and to raise a weight which drives the core barrel into the sediments. The core barrel is plastic and the system is (, suitable for obtaining short cores in soft sediments. Although the core barrel is around 30cm long, the cores obtained were at most 15cm long. Nine samples were obtairied. Positions were obtained using a Garmin eTrex GPS receiver (no differential corrections applied). r· I' ,-- r A Geophysical Investigation of the Derwent Estuary 96 Chapter 5 Seismic Survey Figure 5.18. The Craic corer. The cores were immediately sub-sampled using a 50mL syringe with the end removed, and the wet sediment placed into modified borosilicate glass sample bottles. The bottles had holes drilled into the lids, with Suba seals glued into the holes using a silicon based adhesive. The adhesive was only applied to the outside of the lid, and allowed to cure for several weeks before the bottles were used. These modifications were made so the headspace gas could be extracted (by piercing the seal with a syringe). Since the samples needed to be heated prior to analysis, sample vessel design and material choice was important for several reasons: firstly, we needed to be able to heat the sample vessels without losing gas tightness; and secondly, we needed to be confident the materials used would not evolve light hydrocarbons on heating. A heating and cooling cycle prior to sample collection was performed as a precaution. Around 30mL of wet sediment was retained from each core, and the bottles have lOOmL capacity. A Geophysical Investigation of the Derwent Estuary 97 Chapter 5 Seismic Survey ,:: , -, Figure 5. J9. Modified borosilicate glass sample bottles. The rubber Suba seal inserts in the lids are designed to allow the removal ofgas with a syringe. Typically, headspace gas analysis is perfonned using a flame-ionisation detector (AD) coupled to a gas chromatograph (GC). In this case, a AD was not available so a high resolution mass spectrometer (MS) was used instead. The GC-MS used was the Kratos Analytical Concept ISQ in the Central Science Laboratory. Analyses were kindly perfonned by Or. Noel Oavies. All samples were analysed for methane, and calibrated against authentic standards so results could be reported in parts-per-million (ppm). The results are semi-quantitative at best, since the volume of wet sediment obtained was not consistent and cannot be easily quantified. Several samples were also analysed for hydrogen sulfide, but it was not detected (note that H2S may have been present below detection limits). All except one sample contained elevated (with respect to a blank: and fresh air) concentrations of methane. Results are summarised in table 2 below. Sample Easting Northing Mass Concentration (m) (m) (g) (ppm) 1 529451 5246778 47.03 7 2 530048 5246987 13.42 11 3 530546 5246978 33.50 27 4 531050 5247073 36.83 52 5 531594 5247249 37.36 29 6 532102 5247469 49.30 106 7 532592 5247729 36.84 15 8 533064 5247862 37.77 10 A Geophysical Investigation of the Derwent Estuary 98 Chapter 5 Seismic Survey 1 9 1533213152478491 37.13 1 10 I Table 2. Sample 1 is the only sample that gave an ambient methane concentration - all other samples have elevated concentrations of methane. Although the concentrations of samples 2, 8 and 9 are close to that of sample 1, Dr. Noel Davies considered them significantly different (pers comm 2(01). Since sample 6 had by far the highest methane concentration, it (along with sample 4) was analysed for hydrogen sulfide, but no hydrogen sulfide was detected. Sample 1 contained a high proportion of sand and shell material. Sample 9 was partly sandy and partly muddy, and the other samples were of uniform composition - dark grey mud. 5248000 + 1-5248000 5246000 + 1-5246000 5244000 + 528000 530000 532000 534000 536000 Figure 5.20. Location map of sediment samples showing relative concentrations of methane (normalised by sample mass). Larger dots correspond to higher measured gas concentrations. Note that even after normalisation the samples with the highest concentrations occur in the centre of the Estuary. The yellow region marks the distribution ofacoustic turbidity. The results tend to support the seismic evidence - an elevated concentration of methane is present in the shallow sediments of the Derwent Estuary. Relatively speaking, the highest concentrations (even when scaled for sample mass) occur in the central portion of the Estuary, with reduced concentrations near the riverbanks (this is in accordance with the acoustic turbidity distribution). Although it is impossible to generalise for the entire Estuary based on a single transect of the river, this sampling reinforces the seismic evidence for shallow gas in the Derwent Estuary. A Geophysical Investigation of the Derwent Estuary 99 Chapter 5 Seismic Survey Other Seismic Features Because the seismic system had an analogue recording and printing process, typical seismic post-acquisition processing could not be performed. This precluded the use of techniques such as automatic gain control filtering or automated migration. While manual migration could theoretically be done, it is impractical to migrate even a small section of (-., the data. One ofthe great advantages ofautomated migration is the suppression ofmultiple reflections. There are a number of multiple types in the Derwent Estuary data, )Jut particularly problematic are the almost universal seafloor multiples. Seafloor multiples are characterised by occurring at exact two-way time multiples of the seafloor reflector. They are common because of the generally large acoustic impedance contrast between water and the seafloor and the large contrast between water and air. These impedance contrasts cause the energy to reverberate in the water column - up to four or five sea floor multiples may occur given the right conditions, but at least one is almost always present. Other multiple types that occur on the Derwent Estuary profiles are ordinary multiples from geological interfaces, and "peg-leg" multiples. These concepts are illustrated below. Interface 2 Prim\ Shot r Water () Interface 1 Sand () Interface 2 Bedrock (") Interface 3 Figure 5.21. Schematic ofsome primary reflection travelpaths. Note that although this schematic is shown with non-vertical incidence for clarity (and therefore assuming a significant source-receiver offset), the source-receiver offset for the boomer system was small - so the actual seismic travelpaths were largely vertical. n A Geophysical Investigation of the Derwent EslUary 100 Chapter 5 Seismic Survey Intetface 2 Seafloor Intetface 2 Prim\ /MultiPI, /MultiP!, Shot Water Intetface 1 Sand Intetface 2 Bedrock Intetface 3 Figure 5.22. Schematic ofsome simple multiple reflection travelpaths. Although only one multiple is shown per reflector, it is entirely possible to have several multiples from the same interface. These multiples are common on my reflection profiles. The seafloor multiples occur because ofthe large acoustic impendence contrasts between water and air, and water and sediment. Figure 5.23. This is a simple example of sea floor multiples (although there are other multiples present). Apart from the primary reflectors near the top of this section, most of the other reflections are multiples. There are two strong (and one faint) sea floor multiples in this section. A Geophysical Investigation of the Delwent EstualY 101 Chapter 5 Seismic Survey , Interface 2 Simple Complex Primary Pegleg Pegleg 1 Shot \ ~ / ] r. 11 \ A J I I I I Water I ,. I 11 \ / \ / I I I , I V I I J~ ~ I \ 7 I I II Sand ~ ~ J \ I I I I \/ Interface 2 Bedrock r Interface 3 Figure 5.24. Schematic of some more complicated (pegleg) multiples. In these examples, the energy reverberates within layers of sediment (the Complex peglegs), or travels within a layer of sediment and reverberates in the water column. "Complex Pegleg 1" is shown as a dashed line for clarity only. These kinds ofmultiples also occur on the Derwent Estuary seismic reflection profiles. r •~ r~ A Geophysical Investigation of the Derwent Estuary 102 Chapter 5 Seismic Survey • In ", >i ...... '"'f ~,,-' ~ ... 'It,. ~. ; ... ~ ~ ...... ~~,~ ... :. .... --". .. -,~--~--1---r-----. "f';/1 Figure 5.25. This is an example ofa peg-leg multiple that looks like an ordinary multiple. Beneath the broad zone of acoustic turbidity, there are two multiples (marked A and B). While these appear to be simply the first multiples ofthe seajloor (A) and the zone ofacoustic turbidity (B), they are actually the first multiple of the acoustic turbidity (A) and a pegleg where the energy has reverberated once between the top of the acoustic turbidity and the seajloor (B). This can be demonstrated by looking at the two way times of the multiples with reference to the zero time line. Ifmultiple A was in fact the sea jloor multiple, it would occur at exactly twice the two way time ofthe seajloor. However, it actually occurs at a slightly greater two way time (ie. twice the two way time ofthe top ofthe acoustic turbidity). The seajloor multiple in this case is very faint, and lies just above multiple A. A Geophysical Investigation of the Derwent Estuary 103 Chapter 5 Seismic SUlvey ...... , C'l ,.'t\. "" Figure 5.26. In this example of a pegleg multiple, the energy has travelled to and from reflector A and reverberated once in the water column. This means that the pegleg occurs at the two way time ofreflector A plus the two way time ofthe water. A Geophysical Investigation of the Derwent EsnJalY 104 Chapter 5 Seismic Survey Figure 5.27. This is an example ofa multiple interfering with the interpretation ofprimary reflections. The sea floor multiple is indicated by the positions ofthe arrows, and in the positions marked as A and B, the multiple occurs at the same position as primary reflections (and with similar strength and dip). A Geophysical Investigation of the Derwent Estuary 105 Chapter 6 Acoustic Finite Differencing CHAPTER 6 ACOUSTIC FINITE DIFFERENCING As previously mentioned, geological cross sections are available for the alignments of both the Tasman (Carey and Spry, 1961; also Trollope et aI, 1965) and Bowen Bridges (Colhoun and Moon, 1984). The later Tasman Bridge section is slightly more detailed. The availability of these cross sections provided the possibility of conducting acoustic finite difference modeling. Although I have seismic profiles from both alignments, the Tasman Bridge is more 'interesting' geologically, so it was chosen to be modelled. The section from Carey and Spry (1962) was used to construct the digital representation of the geology used for the models discussed below, although the section from Trollope et al. (1965) was also digitised (there were some difficulties getting the modeling program to accept the more complex section, so it was abandoned). Figure 6.1 Part ofthe Tasman Bridge seismic profile, approximately corresponding to the modelled section. The seafloor profile is slightly different to the published cross sections, because ofthe Lake Illawarra. The Lake Illawarra was a bulk carrier that ran into pylon 17 ofthe bridge and forms the a lump in the deepest part ofthe water. The reflections in the water column are side echoes offthe bridge pylons. A Geophysical Investigation of the DeIWent Estuary 106 Chapter 6 Acoustic Finite Differencing The finite difference modeling was undertaken to attempt to answer two main questions. Firstly I wanted to know if a low velocity, low density gas charged layer of sediment could produce multiple reflections from its upper surface. Secondly, I had observed that no r reflectors were resolvable beneath the acoustically turbid zone on the seismic profile, but the geological sections indicated that there were a number of potential reflectors present (ie. boundaries between various sediment and rock types). I therefore wanted to know if r those reflectors would be resolvable on the Tasman Bridge seismic profile if the silt was not highly attenuating. The acoustic finite difference modeling was carried out using the sufdmod2 routine of Seismic Unix, a freeware seismic processing and modeling package developed by the Centre for Wave Phenomena at the Colorado School of Mines. This routine uses the traditional explicit second order differencing method. The input is a uniformly sampled r two-dimensional velocity model (generated using the unif2 routine), and parameters such as the source .position and runtime duration can be specified. The timestep is not user .specified - the sufdmod2 routine automatically chooses a stable timestep. r· As mentioned, the velocity/density model used by sufdmod2 is produced from the unif2 routine. The input for unif2 is an ASCII file containing x,z coordinate pairs defining points on boundaries between media. The boundaries are then constructed as a series of straight () lines joining the coordinates as defined in the ASCII file. No boundary can contain more than 200 points, and the upper surface of the model must be flat. In my case, the model was constructed based on actual depths reported in the published Tasman Bridge geological section. The model was 766m long, and around 120m deep. The velocities and ("') densities used fell within known ranges for the materials in question (see table I), but without petrophysical data directly from the core logs the values used should be regarded as best estimates. ("') ("') ("') A Geophysicallnvestigation ofthe Denvent Estuary 107 Chapter 6 Acoustic Finite Differencing Dist.~rll~e o 20 40 4500 3750 50 3000 80 2250 100 1500 Velocit' 140 1bO 180 Depth Figure 6.2. The input velocity model for the acoustic finite dijferencing. Although the velocities appear to increase monotonically with depth, I assigned a velocity of l050m/s to the silt when building this model with unij2 (ie. less than the water, and less than it appears to have in this image). The unij2 documentation does not indicate that velocities must increase with depth, but I assume that it must be a 'hidden' requirement. The 'distance' label on the top axis is not physical distance (the model is 766m long) - it is scaled automatically by the display subroutine (ximage). A Geophysical Investigation of the Derwent Estuary 108 Chapter 6 Acoustic Finite Differencing Ig~- - -,~.,-~. ~ xwigl5 . -----Jr1d1J TraGe Nu~lbE;r· 20 40 60 €I0 100 120 140 o 0.2 0,4 _. 0.8 1,2 1.4 1.5 I 1.8 I I Ill: 2.2 ll~ l~ 2,4 IT.me (",eel If"JO l.jj~:1 1 . Hi g h I Frequency Tasman Bridge Model Figure 6.3. The pseduo-high frequency model (see text). Note the scale difference on the vertical axis compared to the other models. This model has some unexpected features' - such as the noise at early times near both ends, and the low frequency event at around Is towards the right of the profile. Since the model was calculated at 5m steps along a 766m input, there are 153 traces. To obtain distance, multiply trace number by 5. A Geophysical Investigation of the Delwent Estuary 109 Chapter 6 Acoustic Finite Differencing o 0.('2 0.')4 0.06 I~I"'-"l,. ... 0.08 0.1 0.12 0.14 0.113 r_--' 0.18 i' '"\ 0.2 wo Wa~ Time (8eCI Tasman Bridge Model - _. - ~.. - --- .._..... Figure 6.4 The first model - the maximum frequency here is around 500Hz. Several multiples occur on this record, and the upper multiple obscures the primary reflections to some extent. However, some reflections can clearly be seen, and the base ofthe silt appears to be resolved. A Geophysical Investigation of the DeIWent Estuary 110 Chapter 6 Acoustic Finite Differencing Tr aGe NI.l~lber 20 40 80 100 120 lAO 0 0.02 0,04 0.06. 0.1 0.12 0.18 0.2 f"_••••, ,'" i,"o 1,la') Tiw" (sec) Low Frequency Tasman Bridge Model Figure 6.5. This is a low frequency model (peak frequency of around 50Hz). Note that the seajloor is not resolved, and there are what appear to be several multiples. This model suggests that conventional oil-field ,-, seismic frequencies would be inappropriate for investigating the detail ofthe Estuary. The sufdmod2 routine is able to calculate source point seismograms, and/or the responses from lines of vertical or horizontal seismograms. The geometry of the boomer seismic f' system is such that the output is effectively a series of closely spaced source point seismograms, since the source-receiver offset is generally small compared to the water depth, and the receiver array is short (several metres only). To effectively simulate the ,-~ A Geophysical Investigation of the Derwent Estuary 111 Chapter 6 Acoustic Finite Differeneing r boomer record, a series of source point seismograms were calculated at 5m increments along the model, starting at one end and "stepping" the source and receiver along together. A shell script was constructed to automate this process (see appendix). Once calculation of r the individual source point seismograms was completed, the output files were concatenated in order of their positions on the model - creating a simulated boomer profile. " After the first high frequency model was calculated, it became apparent that the sufdmod2 routine was not performing the acoustic finite differencing with sufficiently high source frequencies, even though I had instructed it to use very high frequencies. The power r spectra (calculated using the suspecfx routine) revealed the main frequency content was around 500Hz. This is typical of the maximum frequencies recorded for oil field exploration, but is considerably lower than recorded by the boomer system (as mentioned, r' the high pass cut applied during data collection was set at around 5kHz). Although the results were informative, source frequency is inversely related to depth of investigation. This means that by calculating the model with (relatively) low frequencies, the 'virtual boomer' output could potentially contain responses from deep reflectors that the actual r boomer could not have imaged. In an attempt to overcome the frequency limitation of sufdmod2, I scaled my model " vertically and ran it for longer. The scaling applied was a simple multiplication of all interface depths by ten, and the model was calculated for approximately 3.2s (as opposed to 0.25s). I could then display the first 2.5s of data over the same axis length as for the older 0.25s model, effectively compressing the output and (in theory) 'increasing' the [' frequencies by a factor of 10 for display purposes. Unfortunately, the output from this model had some unexpected features and it may be that the simple scaling I used was inappropriate, or there may have been errors in the input parameters. Further work would ('. be required to investigate the problem more thoroughly, but long model run times are limiting - for example, to produce the pseduo-high frequency model took around 13 days of continuous calculation time. In the process, sufdmod2 generated nearly 200 gigabytes of data with around 4.5 million individual timesteps. An even more complicated model was r stopped after around 18 days of runtime when it appeared to be producing meaningless output. .~ A Geophysical Investigation ofthe Derwent Estuary 112 Chapter 6 Acoustic Finite Differencing f' A model was also run at considerably lower frequencies, to investigate the resolution of conventional oil-field style seismic in this setting. As we might expect, the resolution was much lower, and interfaces between thin sediment packages are not resolved. r Conclusions f' The final results revealed that multiple reflections can be generated from the surface ofthe low velocity, low density silt layer. The model also indicates that it should be possible to resolve the base of the silt layer, further reinforcing the suggestion that the silts are attenuating (and probably scattering) the seismic energy. Note that in the accompanying f' model plots, the only gain functions applied are time varying gain and trace amplification with clipping. Although it is possible to apply for complex gain functions to the model output (such as automatic gain control) it would not be appropriate because these gain r functions were not (and could not be) applied to the actual boomer data. (~. (') t'i (') ('j (') A Geophysieal Investigation ofthe Derwent Estuary 113